A multi-lance top-blown continuous converting intelligent system and method

Abstract

The present application relates to the technical field of metal smelting, and particularly relates to a multi-gun top blowing continuous blowing intelligent system and method. The system comprises a data acquisition module, an analysis processing module and a prediction feedback module; the method comprises real-time monitoring of the temperature, liquid level and material state in the furnace, and real-time transmission to a central data processing platform for intelligent feedback control; dimension reduction is performed on the sensing data in the smelting process, and the least principal component with a contribution rate of more than 80% is extracted to realize dimension reduction processing; thermal analysis is performed using material and energy balance equations to construct a theoretical model of the blowing process, capture long-term dependence relationships and predict dynamic parameter changes in the smelting process; an interactive model of the intelligent agent and the smelting environment is constructed to update the control strategy in real time and adaptively adjust the production parameters. The present application improves the low-altitude pollution problem caused by converter blowing, overcomes the low efficiency of current intermittent smelting, and realizes a continuous and stable production process.

Description

Technical Field

[0001] The present invention relates to the technical field of metal smelting, and in particular to an intelligent system and method for multi-lance top-blowing continuous blowing. Background Art

[0002] The production processes for metals such as copper and nickel are primarily categorized into two main categories: pyrometallurgy and hydrometallurgy. Pyrometallurgy dominates, accounting for 86% of the market. Pyrometallurgy involves smelting matte to produce a high-grade metal. After converting, the matte is then electrolytically deposited to produce pure metal. In modern smelters, horizontal converters are commonly used for converting matte. These traditional converter converting processes have limitations, such as cyclical operation, poor thermal stability, and large fluctuations in flue gas concentration, which limit production efficiency and environmental impact. In contrast, the multi-lance continuous top-blowing process successfully addresses these low-level pollution issues and eliminates the safety hazards associated with melt handling. It also offers high efficiency, stable production, and a high degree of automation. With increasing demands for environmental protection and energy efficiency, the multi-lance top-blowing continuous converting process is becoming a new trend in the industry. Although multi-lance top-blown furnace technology has been widely used in the field of copper matte blowing, its application in nickel smelting is still blank. In addition, the existing multi-lance top-blowing blowing technology lacks real-time monitoring of material input, furnace temperature, and molten pool liquid level. This defect weakens the stability and efficiency of the smelting process and affects the service life of the furnace. At present, the adjustment of the spray gun height mainly relies on manual observation. The operator uses the naked eye to judge the position of the spray gun through the observation hole of the furnace body and manually raises and lowers it. At the same time, the uncertainty of the top material input, the air injection amount, and the temperature in the furnace will affect the reaction conditions in the molten pool, which may lead to insufficient or excessive reaction, thereby affecting the metal recovery rate and product quality. This experience-based control method faces high observation difficulty and multiple interference factors in the complex furnace environment, resulting in low overall control accuracy.

[0003] Prior art 1, Chinese patent application number 202411154041.6, discloses an intelligent lance control system and method for supersonic multi-lance top-blowing converting. The system comprises a monitoring module, a feedback control module, a lance lift module, and a lance module. The method involves introducing matte, cold material, solvent, and limestone into a multi-lance top-blowing converting furnace. A temperature monitoring system and a liquid level monitoring system transmit the detected parameters of the converting furnace to an intelligent feedback control module. The feedback control module processes the received data and calculates a height command based on an algorithm. The lance lift module then controls the height of the multi-hole supersonic lances. Oxygen-enriched air flows into the multi-hole supersonic lances, passes through the nozzles, and flows into the multi-lance top-blowing converting furnace. The blister copper is discharged from the blister copper discharge port. While intelligent control and regulation of the entire smelting process is achieved, providing important technical support and significance for improving production efficiency, optimizing product quality, and reducing production costs, issues such as cyclical operation, poor thermal stability, and large fluctuations in flue gas concentration limit production efficiency and environmental impact.

[0004] Prior art 2, Chinese patent, application number 202311689370.6 discloses a multi-gun top-blowing slag reduction device and control method for an oxygen-enriched smelting flash smelting furnace, including a silo and a flash smelting furnace; the silo is filled with a solid reducing agent, and the bottom of the silo is a feeding scraper; the feeding scraper is connected to an inclined chute, and the inclined chute is connected to a straight chute; the straight chute is connected to a buffer bin, and the buffer bin is connected to a pneumatic injection pump; the pneumatic injection pump is connected to a hose, and the spray gun connected to the hose extends into a sedimentation tank, the upper layer of the sedimentation tank is a liquid smelting slag layer, and the lower layer is a liquid copper matte layer, and the end of the spray gun extends into the liquid smelting slag layer. The pneumatic injection pump and spray gun of the present invention work together to transport the solid reducing agent from the silo to the sedimentation tank, where it is evenly mixed with the liquid smelting slag layer, reducing oxides and ferroferric oxide. This improves the flash smelting slag profile, reduces the proportion of copper oxides in the flash smelting slag, ensures that the main form of copper loss is sulfide, improves the flotation efficiency of the slag separation system, reduces the copper content in the tailings, and increases the direct recovery rate and recovery rate of copper in the system. However, there are no strict requirements for the particle size of the raw materials, and the batching work during the raw material preparation process is relatively complicated, increasing the workload during the production process.

[0005] Prior art three, Chinese patent application number 201310243085.1, discloses a dual-furnace, multi-lance, top-blowing continuous converting furnace. This furnace splits the slagging and copper-making converting processes, which occur sequentially in a PS converter, into two converting chambers, achieving a continuous converting process with continuous matte feeding, continuous blasting, continuous slagging, continuous slagging, and continuous or intermittent discharge of blister copper. While replacing the converter's side-blowing with a suspended top-blowing system with multiple lances reduces melt splashing and erosion on furnace bricks, extending their service life, it also allows flux to be injected into the melt through the lance sleeves, ensuring more uniform and rapid contact between the flux and the melt, effectively suppressing the generation and precipitation of large amounts of magnetic iron during the converting process. However, its relatively simple structure and low reducing agent utilization result in high production costs and resource waste.

[0006] Currently, existing technologies 1, 2, and 3 suffer from periodic operation, poor thermal stability, and large fluctuations in flue gas concentration. To address these shortcomings, the present invention proposes an intelligent production process for oxygen-enriched blowing in a multi-lance top-blowing continuous converting furnace. This significant technological innovation builds on traditional PS converter converting technology by intelligently adjusting key operating parameters such as the amount of oxygen-enriched air introduced by the lances, the amount of material added, and the lance height, based on the real-time state of the furnace smelting process. This not only effectively shortens the blowing cycle and improves production efficiency, but also significantly enhances the stability of the blowing process. Summary of the Invention

[0007] The main purpose of the present invention is to provide an intelligent system and method for continuous blowing with multiple lances, so as to solve the problems of periodic operation, poor thermal stability and large fluctuations in flue gas concentration in the prior art.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A multi-lance top-blowing continuous blowing intelligent system, comprising:

[0010] The data acquisition module is used to set up multiple sensors in the furnace to monitor the temperature, liquid level and material status in the furnace in real time, and transmit the collected data to the central data processing platform in real time for intelligent feedback control;

[0011] The analysis and processing module is used to clean various types of data transmitted in real time. It uses principal component analysis to reduce the dimension of sensor data in the smelting process, extracting the minimum principal component with a contribution rate of more than 80% to achieve dimensionality reduction.

[0012] A prediction feedback module is configured to utilize material and energy balance equations to perform thermal analysis, construct a theoretical model of the copper matte converting process, capture long-term dependence, and predict dynamic parameter changes in the smelting process; and construct an interaction model of the agent and the smelting environment, update the control strategy in real time, and adaptively adjust the production parameters.

[0013] As a further improvement of the present application, the data acquisition module comprises:

[0014] A temperature acquisition submodule is configured to arrange temperature sensors at positions such as the furnace wall, the furnace roof and the furnace roof to monitor the overall thermal distribution in the furnace, and install temperature sensors in the material to monitor the material temperature; after obtaining the data, the data acquisition module is configured to transmit the data to the intelligent feedback control system for processing;

[0015] A liquid level acquisition submodule is configured to monitor the liquid level by using a laser liquid level sensor; a part of the laser is transmitted to a time transmitter as a reference signal, and another part is converted into parallel light by an optical system to irradiate the liquid surface of the molten pool, and the signal is amplified by a sampling circuit; after converting the collected electrical signal into a digital signal, the validity of the signal is verified, and the liquid level data is visually displayed;

[0016] A material quantity acquisition submodule is configured to install a camera on the top of the conveyor belt to collect material data on the dynamic conveyor belt in real time, process the collected images by using a filtering algorithm, integrate the processed image data into an analysis processing module, calculate the size and trend of the material flow, and predict the trend of the material flow.

[0017] As a further improvement of the present application, the processing of the intelligent feedback control system in the temperature acquisition submodule comprises:

[0018] A lance height monitoring unit is configured to determine the initial height of the lance in the converting furnace, obtain the image of the lance in the furnace by image acquisition, and analyze the image to obtain the height of the lance from the liquid surface of the molten pool;

[0019] A height correction unit is configured to obtain real-time height data of the liquid surface of the molten pool and lance height data, set safety threshold and effective threshold of the distance difference between the liquid surface of the molten pool and the lance height, and send a safety control signal to the lance control unit if the distance difference between the current liquid surface of the molten pool and the lance height is less than the preset safety threshold; and send an effective control signal to the lance control unit if the distance difference between the current liquid surface of the molten pool and the lance height is greater than the preset safety threshold.

[0020] A lance control unit is configured to receive safety control signals and effective control signals from the height correction unit, control the lifting of the lance according to the type of the received signals, control the lance to rise by an electric or hydraulic method when the safety control signal is received, and control the lance to descend by an electric or hydraulic method when the effective control signal is received.

[0021] As a further improvement of the present invention, the liquid level acquisition submodule includes:

[0022] The reference signal unit is used to emit a beam of near-infrared light using a laser transmitter, wherein a portion of the laser light passes through a semi-transmitting reflector and is input into a time transmitter as a reference signal to form a periodic time sequence;

[0023] The liquid level acquisition unit is used to optically process another part of the laser light to form a parallel beam of set width to illuminate the surface of the molten pool. After part of the laser light is reflected, it is received by the laser level sensor and the received reflected laser light is converted into an electrical signal.

[0024] The information feedback unit is used to amplify the converted electrical signal and align it with the time series. After converting the electrical signal into a digital signal, it verifies the validity of the signal and visualizes the liquid level data. If an unstable liquid level is detected, an alarm is issued and the spray gun is raised to the top.

[0025] As a further improvement of the present invention, the analysis and processing module includes:

[0026] The data preprocessing submodule is used to obtain the furnace temperature, liquid level and material status collected in real time by the data acquisition module, and perform preprocessing such as cleaning, noise processing, outlier detection and elimination, missing value processing, data smoothing and noise reduction, and consistency verification on the acquired real-time data;

[0027] The principal component analysis submodule is used to establish a data matrix under normal operating conditions of a multi-lance top-blown converting furnace and calculate the covariance matrix R. By decomposing the covariance matrix through eigenvalues, the eigenvalues ​​and corresponding eigenvectors are obtained to reflect the correlation between features.

[0028] The importance ranking submodule is used to calculate the importance of each feature and evaluate the impact of the input furnace temperature, liquid level and material state on the smelting quality indicators. By analyzing the cumulative variance contribution rate of the principal components, the minimum principal component with a contribution rate of more than 80% is proposed to achieve dimensionality reduction.

[0029] As a further improvement of the present invention, the prediction feedback module includes:

[0030] The theoretical model submodule is used to construct a theoretical model of the metal matte blowing process. By inputting parameters such as ore ratio and oxygen concentration, it predicts the changing trend of material flow and energy flow and the smelting efficiency.

[0031] The result prediction submodule is used to automatically generate feedback information based on the prediction results of the theoretical model of the metal matte blowing process to inform the operator of possible quality fluctuations or changes in storage conditions;

[0032] The interaction model submodule is used to build an interaction model between the intelligent agent and the smelting environment, define the key parameter states, adjust process parameter actions, and optimize the rewards for smelting quality, energy consumption, and emissions. By updating the control strategy through real-time data, the intelligent agent adaptively adjusts production parameters.

[0033] As a further improvement of the present invention, the theoretical model submodule includes:

[0034] The model building unit is used to obtain historical data of the lances of a multi-lance top-blowing converting furnace, perform thermal analysis using material and energy balance equations, use recurrent neural networks to process time series data, establish a theoretical model of the converting process, and use the collected real-time furnace temperature, liquid level, and material status information as a training set to train the model;

[0035] The model evaluation unit is used to set the inertia weight, learning factor, number of particles and other parameters of the particle swarm optimization method, perform model training on each particle and test its performance indicators, mark the particle with the highest accuracy as the global optimal particle, find the particle with the minimum fitness based on the particle state, and update the model's hyperparameter combination;

[0036] The trend prediction unit is used to input parameters such as ore proportion and oxygen concentration into the theoretical model of the blowing process, predict the changing trend of material flow and energy flow and smelting efficiency, and output the prediction results to the information feedback center.

[0037] As a further improvement of the present invention, the interaction model submodule includes:

[0038] The state definition unit is used to obtain the characteristic value parameters of the furnace temperature, liquid level and material state, define the state as the key characteristic value parameter, define the action as adjusting the process parameters, and define the reward as optimizing the smelting quality, energy consumption and emissions;

[0039] The interactive model unit is used to initialize the agent's neural network weight parameters, detect the current working state, select an action based on the current working state, and provide feedback to the agent on the results of the action. Based on the prediction results of the theoretical model of the metal matte blowing process, the agent updates its strategy and value function to achieve the optimal strategy;

[0040] The adaptive adjustment unit is used to obtain the real-time monitored furnace temperature, liquid level and material status. The intelligent agent dynamically adjusts the smelting process parameters according to the real-time status and visualizes the adaptive decision with an explainable tool.

[0041] To achieve the above object, the present invention also provides the following technical solutions:

[0042] An intelligent method for multi-lance top-blowing continuous blowing, comprising:

[0043] Multiple sensors are installed in the furnace to monitor the temperature, liquid level and material status in real time, and the collected data is transmitted to the central data processing platform in real time for intelligent feedback control;

[0044] Clean all kinds of data transmitted in real time, use principal component analysis to reduce the dimension of sensor data in the smelting process, extract the minimum principal component with a contribution rate of more than 80%, and achieve dimensionality reduction processing;

[0045] Use material and energy balance equations to conduct thermal analysis, construct a theoretical model of the blowing process, capture long-term dependencies, and predict dynamic parameter changes during the smelting process; build an interaction model between the intelligent agent and the smelting environment, update the control strategy in real time, and adaptively adjust production parameters.

[0046] To achieve the above object, the present invention also provides the following technical solutions:

[0047] An intelligent multi-gun top-blowing continuous refining device includes spray guns, wherein the spray guns are installed on the top of the multi-gun top-blowing furnace, with a total of 8 spray guns; a cold material inlet is opened on the furnace wall on the side close to the low-grade metal matte inlet, and the low-grade metal matte inlet is opened obliquely upward and outward; an outlet for refined metallic sulfur is opened obliquely downward and outward on the other side of the inlet; a slag outlet is opened on the furnace wall on the side close to the outlet, and the top of the slag outlet is an ascending flue.

[0048] The present invention adopts a new multi-gun top-blowing furnace to replace the traditional PS converter blowing. This technological innovation not only significantly improves the serious low-altitude pollution problem caused by converter blowing, but also overcomes the current disadvantage of low efficiency of intermittent operation in metal smelting, realizes a continuous and stable production process, and greatly improves the production efficiency of metal smelting. In addition, the application of the new multi-gun top-blowing furnace also optimizes the flue gas treatment system, reduces the complexity and investment cost of the system, and makes the flue gas treatment more efficient. It is conducive to promoting the transformation and upgrading of my country's smelting industry and the construction of ecological civilization; the use of a multi-gun top-blowing intelligent control system can achieve energy conservation and emission reduction, reduce labor costs, enhance safety, extend equipment life, and promote industrial upgrading. At the same time, it brings market competitive advantages to enterprises and supports data-driven decision-making, which is a key measure to promote industrial modernization and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a functional module diagram of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0050] Figure 2 This is a schematic diagram of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0051] Figure 3 This is a functional module diagram of a data acquisition module of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0052] Figure 4 This is a functional module diagram of a temperature acquisition submodule of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0053] Figure 5 This is a functional module diagram of a liquid level acquisition submodule of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0054] Figure 6 This is a functional module diagram of an analysis and processing module of an embodiment of an intelligent system for multi-lance top-blowing continuous blowing according to the present invention;

[0055] Figure 7 This is a functional module diagram of a prediction and feedback module of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0056] Figure 8 This is a functional module diagram of a theoretical model submodule of an embodiment of a multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0057] Figure 9 This is a functional module diagram of an interactive model submodule of an embodiment of the multi-lance top-blowing continuous blowing intelligent system of the present invention;

[0058] Figure 10 This is a schematic diagram of the steps of an embodiment of the intelligent method for multi-lance top-blowing continuous blowing according to the present invention;

[0059] Figure 11 This is a schematic diagram of an embodiment of the intelligent method for continuous blowing with multiple lances and top blowing according to the present invention;

[0060] Figure 12 This is a schematic structural diagram of an embodiment of the multi-lance top-blowing continuous blowing intelligent device of the present invention;

[0061] Figure 13 This is a schematic structural diagram of an embodiment of an electronic device of the present invention;

[0062] Figure 14 This is a schematic structural diagram of an embodiment of a storage medium of the present invention. DETAILED DESCRIPTION

[0063] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0064] The terms "first", "second" and "third" in the present invention are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" and "third" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "multiple" is at least two, for example, two, three, etc., unless otherwise clearly and specifically defined. All directional indications in the embodiments of the present invention (such as up, down, left, right, front, back...) are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0065] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0066] like Figure 1 As shown, this embodiment provides an embodiment of a multi-lance top-blowing continuous blowing intelligent system. In this embodiment, the multi-lance top-blowing continuous blowing intelligent system specifically includes:

[0067] Data acquisition module 1 is used to set up multiple sensors in the furnace to monitor the temperature, liquid level and material status in the furnace in real time, and transmit the collected data to the central data processing platform in real time for intelligent feedback control;

[0068] Analysis and processing module 2 is used to clean various types of data transmitted in real time, use principal component analysis to reduce the dimension of sensor data in the smelting process, extract the minimum principal component with a contribution rate of more than 80%, and achieve dimensionality reduction processing;

[0069] Prediction and feedback module 3 is used to perform thermal analysis using material and energy balance equations, build a theoretical model of the blowing process, capture long-term dependencies, and predict dynamic parameter changes during the smelting process; it also builds an interaction model between the intelligent agent and the smelting environment, updates the control strategy in real time, and adaptively adjusts production parameters.

[0070] Preferably, the data acquisition module 1 of this embodiment can obtain key information such as the temperature, liquid level and material status in the furnace in real time by setting up multiple sensors in the furnace; the collected data is transmitted to the central data processing platform in real time to ensure the timeliness and accuracy of the data; the analysis and processing module 2 removes noise and outliers to improve data quality; through dimensionality reduction processing, the minimum principal component with a contribution rate of more than 80% to the smelting process is extracted, which simplifies the data model and reduces the computational complexity; the prediction and feedback module 3 uses the material and energy balance equation to construct a theoretical model to capture the long-term dependency relationship in the smelting process; predicts the dynamic parameter changes in the smelting process to provide a scientific basis for production adjustment; constructs an interaction model between the intelligent agent and the smelting environment to realize real-time update and self-adjustment of the control strategy (for specific schematic diagrams, please refer to the attached Figure 2 ).

[0071] In summary, the data acquisition module 1 of this embodiment monitors the furnace status in real time, which helps to timely discover and solve potential problems and avoid production interruptions; it provides a rich data basis for intelligent analysis and control, which helps to realize the automation and intelligence of the smelting process; the analysis and processing module 2 improves the efficiency and accuracy of data analysis, and provides more reliable data support for thermal analysis and prediction; the dimensionality reduction processing reduces the amount of calculation, which helps to realize real-time online analysis; the prediction feedback module 3 can more accurately predict the parameter changes in the smelting process through theoretical models and prediction models, and provide strong support for production optimization; the intelligent interaction model enables the smelting process to be adaptively adjusted according to actual conditions, thereby improving the stability and flexibility of the production process; through precise control and prediction, the smelting process can be optimized, energy consumption and production costs can be reduced, and economic benefits can be improved.

[0072] Furthermore, if Figure 3 As shown, the data acquisition module 1 specifically includes:

[0073] The temperature acquisition submodule 11 is used to arrange temperature sensors at the furnace wall, furnace roof and other locations to monitor the overall heat distribution in the furnace, and install temperature sensors in the material to monitor the material temperature; after acquiring the data, it is transmitted to the intelligent feedback control system for processing through the data acquisition module;

[0074] The liquid level acquisition submodule 12 is used to monitor the liquid level using a laser liquid level sensor. A portion of the laser light is transmitted to a time transmitter as a reference signal, while the remaining portion is converted into parallel light by an optical system to illuminate the molten pool liquid surface. The signal is then amplified by a sampling circuit. The collected electrical signal is converted into a digital signal to verify the validity of the signal, and the liquid level data is visually displayed.

[0075] The material quantity collection submodule 13 is used to install a camera on the top of the conveyor belt to collect material data on the dynamic conveyor belt in real time, use a filtering algorithm to process the collected images, and integrate the processed image data into the analysis and processing module to calculate the size and trend of the material flow and predict the material flow trend.

[0076] Preferably, the temperature acquisition submodule 11 of this embodiment can fully and accurately obtain the temperature distribution information in the furnace by arranging temperature sensors at the furnace wall, furnace top and other positions; installing the temperature sensor in the boring can reflect the temperature change of the material in real time and accurately, which is very important for controlling the material quality and process; the temperature data is transmitted to the intelligent feedback control system through the data acquisition module to realize real-time processing and analysis of the data, providing a basis for optimizing the process parameters; the liquid level acquisition submodule 12 adopts a laser liquid level sensor with the characteristics of high precision and high sensitivity, which can accurately monitor the liquid level change of the molten pool; through the optical system and parallel light irradiation, as well as the signal amplification of the sampling circuit The system can improve the signal quality, verify the validity of the signal, and ensure the accuracy of the data; convert the collected liquid level data into a digital signal and then display it visually, so that the operator can intuitively understand the liquid level status; the material quantity acquisition submodule 13 installs a camera on the top of the conveyor belt, which can collect material data on the dynamic conveyor belt in real time, including the shape, size, quantity, etc. of the material; the collected image is processed by a filtering algorithm to improve the image quality; the processed image data is integrated into the analysis and processing module to provide a basis for calculating the material flow; through the calculation of the analysis and processing module, the size and trend of the material flow can be accurately predicted, providing a basis for production scheduling and inventory management.

[0077] To sum up, the temperature acquisition submodule 11 of this embodiment improves the accuracy and stability of temperature control in the furnace, optimizes the process, and improves product quality; monitors material temperature in real time to prevent material damage or process failure due to temperature abnormalities; provides accurate data support for the intelligent feedback control system to achieve automated and intelligent production control; the liquid level acquisition submodule 12 accurately monitors the molten pool liquid level to prevent production accidents caused by liquid level abnormalities; improves the accuracy and stability of liquid level control and optimizes the smelting process; data visualization helps operators respond and make decisions quickly, and improves production efficiency and safety; the material quantity acquisition submodule 13 grasps material flow information in real time and accurately, improves the flexibility and accuracy of production scheduling; predicts material flow trends, helps optimize inventory management, reduces inventory backlogs and waste, improves production efficiency, reduces production costs, and enhances the competitiveness of the enterprise.

[0078] Furthermore, if Figure 4 As shown, the processing performed by the intelligent feedback control system in the temperature acquisition submodule 11 specifically includes:

[0079] The spray gun height monitoring unit 111 is used to determine the initial height of the spray gun in the refining furnace, obtain an image of the spray gun in the furnace through image acquisition, and analyze the image to obtain the height of the spray gun from the molten pool liquid level;

[0080] The height correction unit 112 is used to obtain the real-time height data of the molten pool liquid level and the spray gun height data, set the safety threshold and the effective threshold of the distance difference between the molten pool liquid level and the spray gun height, and send a safety control signal to the spray gun control unit if the current distance difference between the molten pool liquid level and the spray gun height is less than the preset safety threshold; if the current distance difference between the molten pool liquid level and the spray gun height is greater than the preset safety threshold, send an effective control signal to the spray gun control unit;

[0081] The spray gun control unit 113 is used to receive the safety control signal and the effective control signal sent by the height correction unit, and control the raising and lowering of the spray gun according to the type of signal received. When the safety control signal is received, the spray gun is controlled to rise by electric or hydraulic means; when the effective control signal is received, the spray gun is controlled to descend by electric or hydraulic means.

[0082] Preferably, the spray gun height detection unit 111 of this embodiment determines the initial height of the spray gun in the refining furnace, and obtains image information of the spray gun in the furnace in real time through image acquisition technology; analyzes and processes the image to accurately calculate the height of the spray gun from the molten pool liquid surface; the height correction unit 112 obtains the real-time height data of the molten pool liquid surface and the height data of the spray gun; sets a safety threshold and an effective threshold of the difference in distance between the molten pool liquid surface and the spray gun height to determine whether the current state is safe or effective; based on the judgment result, sends a response control signal to the spray gun control unit; the spray gun control unit 113 receives the control signal sent by the height correction unit, and controls the lifting and lowering of the spray gun according to the signal type; realizes precise control of the spray gun through electric or hydraulic means to ensure that the spray gun can move according to the predetermined trajectory and speed.

[0083] In summary, the spray gun height monitoring unit 111 of this embodiment realizes real-time monitoring of the spray gun height, thereby improving the accuracy and safety of the blowing process; through image analysis technology, it avoids the errors and limitations of traditional measurement methods, and improves the accuracy and reliability of measurement; the height correction unit 112 ensures a safe distance between the spray gun and the liquid surface of the molten pool through real-time monitoring and threshold judgment, and prevents safety problems or reduced production efficiency caused by a distance that is too close or too far; it realizes dynamic adjustment of the spray gun height, thereby improving the stability and flexibility of the blowing process; the spray gun control unit 113 realizes precise control of the spray gun height, thereby ensuring the stability and efficiency of the blowing process; through electric or hydraulic means, it improves the reliability and response speed of the spray gun control, providing strong support for the automation and intelligence of the blowing process.

[0084] Furthermore, if Figure 5As shown, the liquid level acquisition sub-module 12 specifically includes:

[0085] a reference signal unit 121 for emitting a beam of near-infrared light using a laser emitter, wherein a portion of the laser light passes through a semi-transmissive mirror and is input into a time transmitter as a reference reference signal to form a periodic time sequence;

[0086] a liquid level acquisition unit 122 for passing another portion of the laser light through optical processing to form a parallel light beam of a set width to irradiate the molten pool surface, and receiving the reflected laser light by a laser liquid level sensor and converting the received reflected laser light into an electrical signal;

[0087] an information feedback unit 123 for amplifying the converted electrical signal and aligning it with the time sequence, verifying the effectiveness of the digital signal after converting the electrical signal into a digital signal, and visually displaying the liquid level data, issuing an alarm if an unstable liquid level is detected, and raising the lance to the topmost end.

[0088] Preferably, the reference signal unit 121 of the present embodiment emits a beam of near-infrared light using a laser emitter to ensure the stability and accuracy of the signal; through a semi-transmissive mirror, a portion of the laser light is input into a time transmitter as a reference reference signal to form a periodic time sequence, providing a time reference for signal processing; the liquid level acquisition unit 122 passes another portion of the laser light through optical processing to form a parallel light beam of a set width to irradiate the molten pool surface, ensuring that the light beam can uniformly irradiate the liquid surface, thereby obtaining more accurate liquid level information; the reflected laser light is received by a laser liquid level sensor and converted into an electrical signal; the information feedback unit 123 amplifies the converted electrical signal and aligns it with the time sequence, ensuring the stability and accuracy of the signal during processing; the digital signal is converted into a digital signal to verify the effectiveness of the signal, which further improves the reliability and accuracy of the data; the liquid level data is visually displayed to facilitate real-time monitoring and recording by the operator; if an unstable liquid level is detected, an alarm is issued and the lance is raised to the topmost end, ensuring the safety and stability of the system.

[0089] To sum up, the reference signal unit 121 of this embodiment ensures the time synchronization of the system, making the collection and processing of liquid level data more accurate and reliable; provides a basis for signal alignment and verification, and improves the overall performance of the system; the liquid level acquisition unit 122 realizes non-contact measurement of the molten pool liquid level, avoiding the errors and pollution caused by traditional measurement methods; improves the accuracy and real-time performance of liquid level measurement, and provides an important basis for industrial production and control; the information feedback unit 123 realizes real-time processing and feedback of liquid level data, and improves the response speed and degree of automation of the system; through visual display and alarm functions, it improves the monitoring efficiency and safety of operators; and takes timely measures when the liquid level is unstable to avoid possible production accidents and equipment damage.

[0090] Furthermore, if Figure 6 As shown, the analysis and processing module 2 specifically includes:

[0091] The data preprocessing submodule 21 is used to obtain the furnace temperature, liquid level and material status collected in real time by the data acquisition module, and perform preprocessing such as cleaning, noise processing, outlier detection and elimination, missing value processing, data smoothing and noise reduction, and consistency verification on the acquired real-time data;

[0092] The principal component analysis submodule 22 is used to establish a data matrix under normal operating conditions of the multi-lance top-blowing converting furnace and calculate the covariance matrix R. By decomposing the covariance matrix by eigenvalues, the eigenvalues ​​and corresponding eigenvectors are obtained to reflect the correlation between the features.

[0093] The importance ranking submodule 23 is used to calculate the importance of each feature and evaluate the impact of the input furnace temperature, liquid level and material state on the smelting quality indicators. By analyzing the cumulative variance contribution rate of the principal components, the minimum principal component with a contribution rate of more than 80% is proposed to achieve dimensionality reduction processing.

[0094] Among them, the data preprocessing submodule 21 adopts dynamic adaptive data cleaning and noise reduction; the dynamic adaptive noise filtering expression is:

[0095]

[0096] Where y i represents the smoothed data point, x i+j represents the original data point, w ij represents a dynamic adaptive weight based on the deviation of the data point from the local mean, φ(x i+j ) represents a nonlinear transformation function, for example, φ(x) = tanh(x), μ i represents the local mean, which is calculated as σ i represents the local standard deviation, calculated as ∈ denotes a small constant to prevent denominator from being zero, for example, ∈ = 10 -6 k denotes the window radius;

[0097] Dynamic outlier detection and removal:

[0098]

[0099] where D i denotes the dynamic Mahalanobis distance, combined with a nonlinear decay factor, X i denotes the i-th data point, μ denotes the data mean vector, ∑ denotes the data covariance matrix, β j denotes the nonlinear decay coefficient, based on feature importance, ψ(X j ) denotes the nonlinear feature transformation function, for example, ψ(X j ) = log(1 + |X j |);

[0100] Dynamic missing value imputation:

[0101]

[0102] where denotes the estimated value of the missing value, β0 denotes the intercept term, β j denotes the regression coefficient, φ(x i,j ) denotes the nonlinear transformation function, μ j denotes the mean of the j-th feature, σ j denotes the standard deviation of the j-th feature;

[0103] The principal component analysis submodule 22 employs multi-level feature decomposition and nonlinear fusion, dynamic covariance matrix calculation:

[0104]

[0105] where R denotes the dynamic covariance matrix, φ(X i ) denotes the nonlinear transformed data point, denotes the nonlinear transformed data mean;

[0106] Dynamic eigenvalue decomposition:

[0107]

[0108] where Q denotes the eigenvector matrix, Λ denotes the eigenvalue diagonal matrix, λ i denotes the nonlinear eigenvalue transformation function;

[0109] Dynamic eigenvector adjustment:

[0110]

[0111] Where, v i represents the dynamically adjusted eigenvector, α ij represents the dynamic weight, v ij represents the nonlinear transformation function, μ v represents the mean of the eigenvector, σ v represents the standard deviation of the eigenvector;

[0112] The importance ranking submodule 23 uses a multi-level feature importance assessment and a dynamic cumulative variance contribution rate:

[0113]

[0114] Where λ' i represents the i-th eigenvalue, k' represents the number of principal components selected, and n' represents the total number of features;

[0115] Dynamic feature importance calculation:

[0116]

[0117] Where, I j Indicates the importance of the j-th feature, α ij represents the dynamic weight, φ(λ i ) represents the nonlinear eigenvalue transformation function, v ij represents the jth component of the i-th eigenvector, μ v represents the mean of the eigenvector, σ v Represents the standard deviation of the feature vector; the dynamic adaptive mechanism introduces dynamic weights and nonlinear transformation functions, so that the model can dynamically adjust according to the data distribution, thereby improving processing efficiency and accuracy; multi-level feature fusion combines nonlinear optimization and multi-level feature decomposition, so that the model can capture data features more comprehensively and is suitable for complex scenarios; self-learning weight adjustment introduces a self-learning weight adjustment mechanism, so that the model can dynamically optimize parameters according to historical data, thereby improving the robustness and adaptability of the model.

[0118] Preferably, the data preprocessing submodule 21 of this embodiment removes irrelevant information or duplicate data in the original data to ensure the accuracy and consistency of the data; reduces random fluctuations or errors in the data to improve the signal-to-noise ratio of the data; identifies and removes data points that do not conform to the expected rules to prevent abnormal data from interfering with subsequent analysis; fills gaps in the data to maintain the integrity of the data; ensures the consistency of data at different time points or from different sources to improve the reliability of the data; the principal component analysis submodule 22 organizes the normal operating condition data of the multi-lance top-blown refining furnace into a matrix form for easy analysis; reveals the correlation between the various dimensions of the data; and decomposes the covariance matrix. , and obtain the eigenvalues ​​and corresponding eigenvectors, which represent the main change trends in the data; the results of the eigenvalues ​​and eigenvectors can reflect the correlation between the various dimensions of the data, which helps to understand the internal structure of the data; the importance ranking submodule 23 evaluates the importance of the features by calculating the degree of influence of each feature on the smelting quality indicators; quantifies the specific impact of input factors such as furnace temperature, liquid level and material state on the smelting quality indicators; by analyzing the cumulative variance contribution rate of the principal components, the principal components with the greatest impact on the smelting quality indicators are determined; the minimum principal component with a contribution rate of more than 80% is proposed to achieve data dimensionality reduction processing and reduce the complexity and computational complexity of subsequent analysis.

[0119] In summary, the data preprocessing submodule 21 of this embodiment can improve the accuracy and efficiency of data analysis through data preprocessing; the preprocessed data is cleaner and more accurate, which helps to reveal the real laws and trends in the data and provide strong support for decision-making; the principal component analysis submodule 22 helps to simplify the data dimension and propose the main changing trends and characteristics in the data; it can understand the data more intuitively and provide support for feature selection and model building; the importance ranking submodule helps to identify the key factors that have the greatest impact on smelting quality indicators, and provide a scientific basis for optimizing the smelting process and improving smelting quality; dimensionality reduction processing can simplify the data model and improve analysis efficiency and accuracy.

[0120] Furthermore, if Figure 7 As shown, the prediction feedback module 3 specifically includes:

[0121] Theoretical model submodule 31 is used to construct a theoretical model of the blowing process. By inputting parameters such as ore ratio and oxygen concentration, it predicts the changing trend of material flow and energy flow and the smelting efficiency.

[0122] The result prediction submodule 32 is used to automatically generate feedback information based on the prediction results of the theoretical model of the blowing process to inform the operator of possible quality fluctuations or changes in storage conditions;

[0123] The interaction model submodule 33 is configured to construct an interaction model of the agent and the smelting environment, define a key parameter state, an adjustment process parameter action, and a reward for optimizing smelting quality, energy consumption, and emissions, update a control strategy in real time, and enable the agent to adaptively adjust a production parameter.

[0124] Preferably, the theoretical model submodule 31 can accurately simulate the physical and chemical reactions in the copper matte converting process by using key parameters such as ore proportioning and oxygen concentration, predict the change trend of material flow, composition change, and energy flow such as input, output, and loss, and provide a scientific basis for optimizing the smelting process; the result prediction submodule 32 can automatically generate feedback information based on the prediction result of the theoretical model, and timely notify the operator of possible quality fluctuations or changes in storage conditions; this helps the operator to quickly respond and adjust the smelting parameters to ensure the stability and safety of the smelting process; the interaction model submodule 33 can define a key parameter state, an adjustment process parameter action, and a reward mechanism for optimizing smelting quality, energy consumption, and emissions, and update a control strategy in real time by constructing an interaction model of the agent and the smelting environment; the agent can adaptively adjust a production parameter to realize intelligent control of the smelting process.

[0125] In summary, the theoretical model submodule 31 of the embodiment improves the controllability and predictability of the smelting process, helps to reduce resource waste, and improves smelting efficiency and product quality; the result prediction submodule 32 enhances the dynamic monitoring and timely adjustment capability of the smelting process, reduces quality risks and safety hazards, and improves overall production efficiency and product quality; the interaction model submodule 33 promotes the automation and intelligence level of the smelting process, improves smelting efficiency and quality, and reduces energy consumption and emissions, which conforms to the development trend of green smelting.

[0126] Further, as shown in Figure 8 the theoretical model submodule 31 specifically includes:

[0127] The model construction unit 311 is configured to acquire historical data of the multiple gun top blowing converter lance, perform thermal analysis by using material and energy balance equations, process time series data by using a recurrent neural network, establish a theoretical model of the converting process, and train the model by using collected real-time furnace temperature, liquid level, and material state information as a training set;

[0128] The model evaluation unit 312 is configured to set parameters such as an inertia weight, a learning factor, and a particle number of a particle swarm optimization method, train and test the performance index of each particle, mark the particle with the highest accuracy as a global optimal particle, find a particle with the minimum fitness according to the particle state, and update the hyperparameter combination of the model;

[0129] The trend prediction unit 313 is used to input parameters such as ore proportion and oxygen concentration into the theoretical model of the blowing process, predict the changing trend of material flow and energy flow and smelting efficiency, and output the prediction results to the information feedback center.

[0130] Preferably, the model building unit 311 of this embodiment collects historical data from the lances of the multi-lance top-blowing refining furnace, and uses the material and energy balance equations to perform thermal analysis, which helps to understand the physical and chemical changes in the refining process; uses a recurrent neural network to process time series data, which can capture the time dependence and trend in the data; uses the real-time furnace temperature, liquid level and material status information as a training set to train the theoretical model to improve the accuracy and adaptability of the model; the model evaluation unit 312 sets various parameters of the particle swarm optimization method, such as inertia weight, learning factor and number of particles, and evaluates each particle. The model is trained and its performance indicators are tested to find the optimal combination of model parameters; the particle with the highest labeling accuracy is marked as the global optimal particle, and the particle with the lowest fitness is found based on the particle state, which helps to further adjust and optimize the model; the model's hyperparameter combination is updated based on the evaluation results to improve the model's performance; the trend prediction unit 313 inputs key parameters such as ore ratio and oxygen concentration into the theoretical model of the blowing process; the changing trend of logistics and energy flow and smelting efficiency are predicted, which helps to understand possible changes in the blowing process in advance; the prediction results are output to the information feedback center so that corresponding control measures can be taken in time.

[0131] In summary, the model construction unit 311 of this embodiment establishes a theoretical model based on actual data that can reflect the blowing process; the RNN is used to process time series data, thereby improving the model's ability to capture dynamic changes in the blowing process; the model evaluation unit 312 optimizes the model through the PSO method, thereby improving the model's accuracy and generalization ability, realizing automatic adjustment and optimization of model parameters, and reducing manual intervention; the trend prediction unit 313 realizes real-time monitoring and prediction of the blowing process, thereby improving production efficiency and safety; and guiding production operations through prediction results helps optimize the smelting process and reduce energy consumption and costs.

[0132] Furthermore, if Figure 9 As shown, the interaction model submodule 33 specifically includes:

[0133] The state definition unit 331 is used to obtain characteristic value parameters of the furnace temperature, liquid level, and material state, define the state as the characteristic value key parameter, define the action as adjusting the process parameters, and define the reward as optimizing the smelting quality, energy consumption, and emissions;

[0134] An interaction model unit 332 is configured to initialize the neural network weight parameters of the agent, detect the current working condition state, select an action according to the current working condition, and feed back the feedback obtained by executing the action to the agent, update the policy and the value function of the agent according to the prediction result of the theoretical model of the blowing process, and achieve the optimal policy.

[0135] The process of updating the policy and the value function of the agent is as follows:

[0136] The complex equation of deep deterministic policy gradient, the Actor network update equation is as follows:

[0137]

[0138] In the formula, θ μ represents the parameters of the Actor network, α μ represents the learning rate of the Actor network, Q(s, a|θ Q ) represents the Q value function of the Critic network, and θ Q represents the parameters of the Critic network, μ(s|θ μ ) represents the policy function of the Actor network, and θ μ represents the parameters of the Actor network, represents the Q value gradient of the action a, represents the policy gradient of the Actor network parameters θ μ ;

[0139] The Critic network update equation is as follows:

[0140]

[0141] In the formula, the target value y t is defined as:

[0142] y t = r t+1 + γQ(s t+1 , μ(s t+1 | θ μ )| θ Q )

[0143] In the formula, y t represents the parameters of the Critic network, α Q represents the learning rate of the Critic network, y t represents the target Q value, which combines the Q value of the immediate reward and the next state, r t+1 represents the immediate reward, γ represents the discount factor, and (s t+1 | θ μ ) represents the policy output of the next state s t+1 of the Actor network.

[0144] The policy update equation for proximal policy optimization is:

[0145]

[0146] Where r t (θ) represents the strategy ratio, represents the advantage function estimate, ∈ represents the clipping range parameter, which is usually 0.1 or 0.2, clip(r t (θ),1-∈,1+∈) represents the clipping function, which limits the update amplitude of the strategy ratio;

[0147] Value function update equation:

[0148]

[0149] Where V θ (s t ) represents the current state s t The value function estimation under Represents the target value function, which usually consists of the immediate reward and the value function of the next state;

[0150] Complex state value function update (combined with multi-step returns):

[0151]

[0152] In the formula, the multi-step return Defined as:

[0153]

[0154] Where, represents a multi-step reward, which combines the reward of the next n steps and the state value after n steps; γ represents the discount factor; (s t+n” ) represents the state value after n steps. The complex equations above demonstrate how, in complex industrial environments, an intelligent agent updates its policy and value function through deep reinforcement learning algorithms (such as DDPG and PPO). These algorithms not only consider immediate rewards but also incorporate future multi-step rewards and policy ratio tailoring to ensure the stability and efficiency of policy updates. Through these complex mathematical models, the intelligent agent is able to adaptively adjust process parameters during complex smelting processes to optimize smelting quality, energy consumption, and emissions.

[0155] The adaptive adjustment unit 333 is used to obtain the real-time monitored furnace temperature, liquid level and material status. The intelligent agent dynamically adjusts the smelting process parameters according to the real-time status and visualizes the adaptive decision with an explainable tool.

[0156] Preferably, the state definition unit 331 of this embodiment extracts key eigenvalue parameters by real-time monitoring of the temperature, liquid level and material state in the furnace; clearly defines the state of the smelting process as these eigenvalue key parameters, and defines the action as adjusting the process parameters; takes optimizing smelting quality, energy consumption and emissions as reward goals, providing a clear direction for the learning of the intelligent agent; the interactive model unit 332 initializes the neural network weight parameters of the intelligent agent, so that the intelligent agent can detect the current working state in real time and select an optimal action based on these states; based on the feedback obtained from executing the action and the prediction results of the theoretical model of the blowing process, the intelligent agent can continuously update its strategy and value function until the optimal strategy is reached; the adaptive adjustment unit 333 can detect the temperature, liquid level and material state in the furnace in real time, providing a basis for the decision-making of the intelligent agent; the intelligent agent can dynamically and adaptively adjust the smelting process parameters according to the real-time state to achieve flexible control of the smelting process; the adaptive decision is visualized through interpretable tools, thereby improving the transparency and credibility of the decision.

[0157] In summary, the state definition unit 331 of this embodiment provides a reliable data basis for intelligent decision-making by accurately defining the state of the smelting process; the setting of the reward mechanism makes the learning process of the intelligent agent more targeted and efficient; the interactive model unit 332 realizes the precise control and optimization of the smelting process through the learning and optimization process of the intelligent agent; improves the automation level and efficiency of the smelting process, and reduces energy consumption and emissions; the adaptive adjustment unit 333 realizes adaptive control of the smelting process, and improves the robustness and stability of the system; the visual display of the decision helps operators understand and trust the decision-making process of the intelligent agent, and enhances the operability and human-computer interaction of the system.

[0158] like Figure 10 As shown, this embodiment further provides an embodiment of a multi-lance top-blowing continuous blowing intelligent method. In this embodiment, the multi-lance top-blowing continuous blowing intelligent method is applied to the multi-lance top-blowing continuous blowing intelligent system in the above embodiment. The multi-lance top-blowing continuous blowing intelligent method includes:

[0159] Step S100: multiple sensors are installed in the furnace to monitor the temperature, liquid level and material status in the furnace in real time, and the collected data is transmitted to the central data processing platform in real time for intelligent feedback control;

[0160] Step S200: Cleaning various types of data transmitted in real time, using principal component analysis to reduce the dimension of the sensor data in the smelting process, extracting the minimum principal component with a contribution rate of more than 80%, and achieving dimensionality reduction processing;

[0161] Step S300: Use material and energy balance equations to conduct thermal analysis, build a theoretical model of the blowing process, capture long-term dependencies, and predict dynamic parameter changes during the smelting process; build an interaction model between the intelligent agent and the smelting environment, update the control strategy in real time, and adaptively adjust production parameters.

[0162] Preferably, in step S100 of this embodiment, by setting multiple sensors in the furnace, key information such as the temperature, liquid level and material status in the furnace can be obtained in real time; the collected data is transmitted to the central data processing platform in real time to ensure the timeliness and accuracy of the data; step S200 removes noise and outliers to improve data quality; through dimensionality reduction processing, the minimum principal component with a contribution rate of more than 80% to the smelting process is extracted, which simplifies the data model and reduces the computational complexity; step S300 uses the material and energy balance equation to construct a theoretical model to capture the long-term dependency relationship in the smelting process; predicts the dynamic parameter changes in the smelting process to provide a scientific basis for production adjustment; constructs an interactive model between the intelligent agent and the smelting environment to realize real-time update and self-adjustment of the control strategy (for specific schematic diagrams, please refer to the attached Figure 11 ).

[0163] In summary, the real-time monitoring of the furnace status in step S100 of this embodiment helps to timely discover and solve potential problems and avoid production interruptions; it provides a rich data basis for intelligent analysis and control, and helps to realize the automation and intelligence of the smelting process; step S200 improves the efficiency and accuracy of data analysis, and provides more reliable data support for thermal analysis and prediction; dimensionality reduction processing reduces the amount of calculation, and helps to realize real-time online analysis; step S300 can more accurately predict parameter changes in the smelting process through theoretical models and predictive models, and provide strong support for production optimization; the intelligent interactive model enables the smelting process to be adaptively adjusted according to actual conditions, thereby improving the stability and flexibility of the production process; through precise control and prediction, the smelting process can be optimized, energy consumption and production costs can be reduced, and economic benefits can be improved.

[0164] like Figure 12 As shown, this embodiment also provides an embodiment of a multi-lance top-blowing continuous blowing intelligent device. In this embodiment, the multi-lance top-blowing continuous blowing intelligent device is applied to the multi-lance top-blowing continuous blowing intelligent system in the above embodiment. The multi-lance top-blowing continuous blowing intelligent device includes: a spray gun 1, a cold material inlet 2, a low-grade metal matte inlet 3, a high-grade metal matte outlet 4, a slag outlet 5, and an ascending flue 6;

[0165] Among them, the spray gun 1 is installed on the top of the multi-gun top-blowing furnace, with a total of 8 guns; the cold material inlet 2 is opened on the furnace wall on the side close to the low-grade metal matte inlet 3, and the low-grade metal matte inlet 3 is opened obliquely upward and outward; the high-grade metal matte outlet 4 is opened obliquely downward and outward on the other side of the inlet 3; the slag outlet 5 is opened on the furnace wall on the side close to the high-grade metal matte outlet 4, and the top of the slag outlet 5 is an ascending flue 6.

[0166] Preferably, in this embodiment, matte (primarily composed of FeS, NiS, CuS, PbS, ZnS, CoS, and FeO) produced through matte smelting in flash furnaces, blast furnaces, electric furnaces, and other smelting equipment flows continuously through troughs into a multi-lance top-blown furnace. Subsequently, air enriched with approximately 37% oxygen is introduced into the furnace via multiple lances positioned at the top of the multi-lance top-blown furnace. Simultaneously, an appropriate amount of quartz is added. The FeS and other impurities in the matte are oxidized to form slag, which is partially oxidized along with other volatile impurities and discharged with the flue gas, resulting in matte (primarily composed of NiS and CuS) with a high valuable metal content and converted slag with a low valuable metal content. These slags are separated into separate layers due to their different densities, with the low-density converted slag floating to the top and being removed. Most of the nickel and copper in the high-grade matte remain in the form of sulfides, with a small portion in the form of alloys. Precious metals and some cobalt also enter the high-grade matte. The high-grade matte is siphoned off from the end of the converting furnace. The heat required for converting the high-grade matte is provided by the heat from the oxidation of iron and sulfur in the low-grade matte and the slagging of ferrous oxide. The converting process is not only self-heating, but also produces excess heat that can be used to treat cold charge. Treating cold charge is a key measure to prevent overheating during the converting process and extend the life of the masonry.

[0167] The difference between the blowing of low-nickel matte and copper matte lies in the fact that copper blowing is divided into two stages: the initial stage is the slagging process in which FeS is oxidized to form white copper matte (Cu2S), and the second stage is the oxidation of Cu2S to obtain blister copper, which are respectively called the slagging period and the copper-making period. The blowing of low-nickel matte only goes through the slagging period, and finally produces high-nickel matte (Ni3S2). At a blowing temperature of 1523K, the oxidation reaction of metal sulfide (MeS) can be expressed as:

[0168]

[0169] MeS+O2=Me+SO2

[0170] In summary, the top of the multi-lance top-blowing furnace in this embodiment is provided with 8 spray guns 1, and the oxygen-enriched air required for blowing is sprayed into the furnace through the spray guns 1; the spray gun 1 is made of high-temperature resistant alloy or corrosion-resistant material, and the lifting device of the spray gun 1 is a detachable motor or cylinder, which is convenient for regular inspection and maintenance; the lifting action of the motor, pneumatic cylinder or hydraulic system is controlled by PLC; the PLC control system can automatically adjust the position of the spray gun according to timely feedback and process requirements; the multi-lance top-blowing furnace is equipped with a cold material inlet 2. In the blowing process of low-grade metallic sulfur, cold material must be added to control the furnace temperature and ensure the stability of operation; cold materials suitable for this process include steamed bun shells, splashes generated by blowing, slag after smelting, high-metal content ores and other substances; the cold material should maintain low moisture, and the moisture content before entering the furnace shall not exceed 4%, and large pieces of material shall be pre-crushed to 5 to 25 mm; the low-grade metallic sulfur inlet 3 is mainly responsible for accurately inputting low-grade metallic sulfur into the multi-lance top Blowing furnace to ensure its uniform distribution and reaction; according to the reaction process in the furnace, accurately control the feeding rate of low-grade metallic sulfur to avoid too fast or too slow feeding speed causing temperature fluctuation or incomplete reaction in the furnace; the high-grade metallic sulfur produced by the multi-lance top-blown furnace is continuously discharged through the siphon port; the discharge port is designed to have a certain inclination angle, using gravity to help flow to avoid blockage; the material of the discharge port needs to have the characteristics of high temperature resistance, corrosion resistance and wear resistance; the slag produced by the multi-lance top-blown refining furnace is discharged into the slag bag through the slag port overflow, and the discharge port design needs to be able to withstand high temperature and high pressure while ensuring the smooth outflow of slag; regular discharge can maintain the continuity of the smelting process and improve oxygen utilization; the structure of the rising flue 6 is vertical and made of refractory bricks; in the continuous air supply operation, the flue gas generated by the low-grade metallic sulfur blowing has a high sulfur dioxide content. This flue gas is mainly used for sulfuric acid production, so it has a high temperature and is recovered through the waste heat boiler; such as Figure 13 As shown, this embodiment provides an embodiment of an electronic device. In this embodiment, the electronic device 7 includes a processor 71 and a memory 72 coupled to the processor 71.

[0171] The memory 72 stores program instructions for implementing the intelligent method for multi-lance top-blowing continuous blowing according to any one of the above embodiments.

[0172] The processor 71 is used to execute program instructions stored in the memory 72 to perform intelligent multi-lance top-blowing continuous blowing.

[0173] The processor 71 may also be referred to as a CPU (Central Processing Unit). The processor 71 may be an integrated circuit chip having signal processing capabilities. The processor 71 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The general-purpose processor may be a microprocessor or any conventional processor.

[0174] Furthermore, Figure 14 This is a schematic diagram of the structure of the storage medium of an embodiment of the present application. The storage medium 8 of the embodiment of the present application stores program instructions 81 that can implement all the above methods, wherein the program instructions 81 can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, server, mobile phone, and tablet.

[0175] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0176] In addition, the functional units in the various embodiments of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware or in the form of software functional units. The above is only an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0177] The above detailed description of the specific embodiments of the invention is intended to be illustrative only, and the present invention is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of the present invention. Therefore, equivalent changes, modifications, and improvements made without departing from the spirit and scope of the present invention are also encompassed within the scope of the present invention.

Claims

1. A multi-lance top-blowing continuous refining intelligent system, characterized in that: The multi-lance top-blowing continuous blowing intelligent system includes: The data acquisition module is used to set up multiple sensors in the furnace to monitor the temperature, liquid level and material status in the furnace in real time, and transmit the collected data to the central data processing platform in real time for intelligent feedback control; The analysis and processing module is used to clean various types of data transmitted in real time and use principal component analysis to reduce the dimension of sensor data in the smelting process to achieve dimensionality reduction processing; The principal component analysis submodule is used to establish a data matrix under normal operating conditions of a multi-lance top-blown converting furnace and calculate the covariance matrix R. By decomposing the covariance matrix through eigenvalues, the eigenvalues ​​and corresponding eigenvectors are obtained to reflect the correlation between features. The importance ranking submodule is used to calculate the importance of each feature and evaluate the impact of the input furnace temperature, liquid level, and material state on the smelting quality indicators. By analyzing the cumulative variance contribution rate of the principal components, the minimum principal component with a contribution rate exceeding 80% is proposed to achieve dimensionality reduction. The prediction and feedback module is used to perform thermal analysis using material and energy balance equations, build a theoretical model of the blowing process, capture long-term dependencies, and predict dynamic parameter changes during the smelting process; it also builds an interaction model between the intelligent agent and the smelting environment, updates the control strategy in real time, and adaptively adjusts production parameters.

2. The multi-lance top-blowing continuous blowing intelligent system according to claim 1 is characterized in that: Data acquisition module, including: The temperature acquisition submodule is used to place temperature sensors on the furnace wall, furnace roof and furnace top to monitor the overall heat distribution in the furnace, and install temperature sensors in the material to monitor the material temperature; after acquiring the data, it is transmitted to the intelligent feedback control system for processing through the data acquisition module; The liquid level acquisition submodule uses a laser level sensor to monitor the liquid level. A portion of the laser light is transmitted to a time transmitter as a reference signal, while the remaining portion is converted into parallel light by an optical system to illuminate the molten pool surface. The signal is then amplified by a sampling circuit. The collected electrical signal is converted into a digital signal to verify the signal's validity, and the liquid level data is visualized. The material quantity collection submodule is used to install a camera on the top of the conveyor belt to collect material data on the dynamic conveyor belt in real time. The collected images are processed using a filtering algorithm. The processed image data is integrated into the analysis and processing module to calculate the size and trend of the material flow and predict the material flow trend.

3. The multi-lance top-blowing continuous blowing intelligent system according to claim 2, characterized in that: Temperature acquisition submodule, including: The spray gun height monitoring unit is used to determine the initial height of the spray gun in the refining furnace, obtain an image of the spray gun in the furnace through image acquisition, and analyze the image to obtain the height of the spray gun from the molten pool liquid level; A height correction unit is used to obtain real-time height data of the molten pool liquid level and the spray gun height data, and to set a safety threshold and an effective threshold for the distance difference between the molten pool liquid level and the spray gun height. If the current distance difference between the molten pool liquid level and the spray gun height is less than the preset safety threshold, a safety control signal is sent to the spray gun control unit; if the current distance difference between the molten pool liquid level and the spray gun height is greater than the preset safety threshold, an effective control signal is sent to the spray gun control unit; The spray gun control unit is used to receive the safety control signal and the effective control signal sent by the height correction unit, and control the lifting and lowering of the spray gun according to the type of signal received. When the safety control signal is received, the spray gun is controlled to rise by electric or hydraulic means; when the effective control signal is received, the spray gun is controlled to fall by electric or hydraulic means.

4. The multi-lance top-blowing continuous blowing intelligent system according to claim 2, characterized in that: Liquid level acquisition submodule, including: The reference signal unit is used to emit a beam of near-infrared light using a laser transmitter, wherein a portion of the laser light passes through a semi-transmitting reflector and is input into a time transmitter as a reference signal to form a periodic time sequence; The liquid level acquisition unit is used to optically process another part of the laser light to form a parallel beam of set width to illuminate the surface of the molten pool. After part of the laser light is reflected, it is received by the laser level sensor and the received reflected laser light is converted into an electrical signal. The information feedback unit is used to amplify the converted electrical signal and align it with the time series. After converting the electrical signal into a digital signal, it verifies the validity of the signal and visualizes the liquid level data. If an unstable liquid level is detected, an alarm is issued and the spray gun is raised to the top.

5. The multi-lance top-blowing continuous blowing intelligent system according to claim 1 is characterized in that: Analysis and processing module, including: The data preprocessing submodule is used to obtain the furnace temperature, liquid level and material status collected in real time by the data acquisition module, and to clean, process noise, detect and eliminate outliers, process missing values, smooth and reduce noise, and perform consistency check preprocessing on the acquired real-time data.

6. The multi-lance top-blowing continuous blowing intelligent system according to claim 1, characterized in that: Prediction feedback module, including: The theoretical model submodule is used to construct a theoretical model for the converting process of metal mattes such as copper and nickel mattes. By inputting ore ratio and oxygen concentration parameters, it predicts the changing trend of material flow and energy flow and smelting efficiency. The result prediction submodule is used to automatically generate feedback information based on the prediction results of the theoretical model of the metal matte blowing process such as copper and nickel matte, notifying operators of possible quality fluctuations or changes in storage conditions; The interaction model submodule is used to build an interaction model between the intelligent agent and the smelting environment, define the key parameter states, adjust process parameter actions, and optimize the rewards for smelting quality, energy consumption, and emissions. By updating the control strategy through real-time data, the intelligent agent adaptively adjusts production parameters.

7. The multi-lance top-blowing continuous blowing intelligent system according to claim 6, characterized in that: Theoretical model submodule, including: The model building unit is used to obtain historical data on the lances of a multi-lance top-blowing converting furnace, perform thermal analysis using material and energy balance equations, and use recurrent neural networks to process time series data. This allows the establishment of a theoretical model for the converting process of metal mattes such as copper and nickel mattes. The model is then trained using the collected real-time furnace temperature, liquid level, and material status information as a training set. The model evaluation unit is used to set the inertia weight, learning factor, and particle number parameters of the particle swarm optimization method, perform model training on each particle and test its performance indicators, mark the particle with the highest accuracy as the global optimal particle, find the particle with the minimum fitness based on the particle state, and update the model's hyperparameter combination; The trend prediction unit is used to input parameters such as ore ratio and oxygen concentration into the theoretical model of the blowing process of metal mattes such as copper and nickel mattes, predict the changing trend of material flow and energy flow and smelting efficiency, and output the prediction results to the information feedback center.

8. The multi-lance top-blowing continuous blowing intelligent system according to claim 6, characterized in that: Interaction model submodule, including: The state definition unit is used to obtain the characteristic value parameters of the furnace temperature, liquid level and material state, define the state as the key characteristic value parameter, define the action as adjusting the process parameters, and define the reward as optimizing the smelting quality, energy consumption and emissions; The interactive model unit is used to initialize the agent's neural network weight parameters, detect the current working state, select an action based on the current working state, and provide feedback to the agent. Based on the prediction results of the theoretical model of the smelting process of metal mattes such as copper and nickel mattes, the agent updates its strategy and value function to achieve the optimal strategy; The adaptive adjustment unit is used to obtain the real-time monitored furnace temperature, liquid level and material status. The intelligent agent dynamically adjusts the smelting process parameters according to the real-time status and visualizes the adaptive decision with an explainable tool.

9. An intelligent method for multi-lance top-blowing continuous blowing, applied to the intelligent system for multi-lance top-blowing continuous blowing according to any one of claims 1 to 8, characterized in that: The multi-lance top-blowing continuous blowing intelligent system includes: Multiple sensors are installed in the furnace to monitor the temperature, liquid level and material status in real time, and the collected data is transmitted to the central data processing platform in real time for intelligent feedback control; Clean all kinds of data transmitted in real time, use principal component analysis to reduce the dimension of sensor data in the smelting process, extract the minimum principal component with a contribution rate of more than 80%, and achieve dimensionality reduction processing; Use material and energy balance equations to conduct thermal analysis, construct a theoretical model of the blowing process of metal mattes such as copper and nickel mattes, capture long-term dependencies, and predict dynamic parameter changes during the smelting process; build an interaction model between the intelligent agent and the smelting environment, update the control strategy in real time, and adaptively adjust production parameters.

10. A multi-lance top-blowing continuous blowing intelligent device, applied to the multi-lance top-blowing continuous blowing intelligent system according to any one of claims 1 to 8, characterized in that: The multi-gun top-blowing continuous refining intelligent device includes a spray gun; the spray gun is installed on the top of the multi-gun top-blowing furnace; the cold material inlet is opened on the furnace wall on the side close to the inlet of low-grade metal mattes such as copper matte and low-nickel matte, and the inlet is opened obliquely upward and outward; the outlet of high-grade metal mattes such as blister copper and high-nickel matte is opened obliquely downward and outward on the other side of the inlet of low-grade metal mattes such as copper matte and low-nickel matte; the slag outlet is opened on the furnace wall on the side close to the outlet of low-grade metal mattes such as copper matte and low-nickel matte, and the top of the slag outlet is an ascending flue.

Images