Self-adaptive porous rotary supersonic injection control method and system for top-blown furnace

Abstract

The invention discloses a top-blown furnace self-adaptive porous rotary supersonic injection control method and system, and particularly relates to the technical field of metallurgical furnace charge conveying and control, and the method comprises the following steps: installing a flow sensor, a pressure sensor, a vertical infrared thermometer and other sensors in a feeding pipeline and an injection system; state signals such as feeding flow, pressure, temperature and injection state which are collected in real time are constructed; the real-time state signal is transmitted to a PLC control system through an analog signal; the PLC control system identifies a fault abnormal condition according to the real-time state signal and automatically adjusts a control strategy; performing control analysis on the fault abnormal data; a molten pool behavior prediction model is constructed based on historical abnormal data, and a controlled state signal is collected; and transmitting the controlled state signal to a molten pool behavior prediction model to predict whether the controlled state signal is qualified or not. According to the invention, high-speed, uniform and stable injection and whole-process self-adaptive adjustment of the furnace charge under complex working conditions are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metallurgical furnace charging and control, and particularly relates to a top-blown furnace self-adaptive multi-hole rotary supersonic injection control method and system. BACKGROUND

[0002] With the development of metallurgical processes, top-blown furnaces have higher requirements for the stirring efficiency of the molten pool, the injection range and the reaction rate. Especially under the trend of efficient smelting and green metallurgy, realizing rapid and uniform mixing of furnace charges and sufficient diffusion of reaction products has become a key problem to improve production efficiency and product quality. The traditional feeding mode mostly uses fixed single-hole or multi-hole nozzles, which have the disadvantages of fixed jet direction, single injection distribution and limited flow regulation mode, and it is difficult to meet the complex and variable working condition requirements in the furnace. On the other hand, the existing injection control system mostly relies on manual setting or simple control loop adjustment, and the response speed is limited. When the furnace condition fluctuates rapidly or abnormal working conditions occur, it is difficult to make accurate control in time. Therefore, a rotatable multi-hole nozzle, supersonic injection combined with real-time sensing and adaptive control feeding system is needed to improve the uniformity, efficiency and stability of material injection to adapt to the increasingly complex furnace conditions.

[0003] Prior art one, Chinese patent, patent number: 202510656547.5 relates to the technical field of tin smelting, and discloses a top-blown supersonic cyclone injection lance, a control system and a method thereof. The device includes four layers of sleeves, wherein the flow channel formed by the outer sleeve and the supersonic sleeve is mainly used for injecting desulfurization concentrate; the flow channel between the supersonic sleeve and the cyclone sleeve is used for injecting oxygen-enriched air; the flow channel between the cyclone sleeve and the inner sleeve is used for injecting dry powder; and the inner flow channel of the inner sleeve is used for injecting pulverized coal. The system includes a Mach number acquisition module that acquires the real-time industrial application of oxygen-enriched air flow, pressure, temperature, and lance structure size parameters and the required final lance expansion flow channel outlet Mach number; a three-dimensional flow channel model module that establishes a three-dimensional flow channel model and divides the three-dimensional flow channel model into hexahedral grids; and a preset deviation value module that calculates the initial physical field. Although the oxygen-enriched air is effectively accelerated to a supersonic state, the molten pool stirring efficiency and material particle penetration depth are significantly improved; however, the temperature control is uneven, the reaction is insufficient, the product composition fluctuates, and the quality decreases.

[0004] The prior art two, Chinese patent, patent number: 202410640579.1 provides a top blowing furnace smelting stability intelligent adjusting system, comprising: a lance immersed in the bottom of the smelting reaction pool of the top blowing furnace, for blowing oxidizing agent and coal powder into the smelting reaction pool; a top blowing furnace; a monitoring system for monitoring real-time lance information of the lance and real-time top blowing furnace information of the top blowing furnace, and transmitting the real-time lance information and the real-time top blowing furnace information to the intelligent adjusting system, the real-time lance information including lance end pressure information and lance coal powder flow information, and the real-time top blowing furnace information including top blowing furnace temperature information and furnace material quantity information; an intelligent adjusting system for comparing the real-time lance information, the real-time top blowing furnace information and their corresponding standard values; although, according to the comparison result, the lance or the top blowing furnace is intelligently controlled, through real-time monitoring of the top blowing furnace information and the lance information, precise intelligent control of the top blowing furnace and the lance is realized, the stability and efficiency of the smelting process are improved, and the labor cost is reduced; however, part of the material particles are not effectively injected into the molten metal and are discharged, causing resource waste.

[0005] The prior art three, Chinese patent, patent number: 202311689370.6 discloses an oxygen-enriched smelting flash smelting furnace multi-lance top blowing slag reduction device and control method, comprising a stock bin and a flash smelting furnace; the stock bin is provided with a solid reducing agent, and the bottom of the stock bin is provided with a feeding scraper; the feeding scraper is connected with an inclined chute, and the inclined chute is communicated with a straight chute; the straight chute is communicated with a buffer bin, and the buffer bin is connected with a pneumatic injection pump; the pneumatic injection pump is connected with a rubber pipe, and the rubber pipe is connected with a lance extending into a sedimentation tank; an upper layer of the sedimentation tank is a liquid smelting slag layer, and a lower layer is a liquid copper matte layer, and a distal end of the lance extends into the liquid smelting slag layer. Although the pneumatic injection pump cooperates with the lance to deliver the solid reducing agent in the stock bin to the sedimentation tank, so that the solid reducing agent is uniformly mixed with the liquid smelting slag layer, the oxide and the magnetite are reduced, the flash smelting furnace slag type is improved, the proportion of copper oxide in the flash smelting slag is reduced, the main loss form of copper is ensured to be mainly sulfide, the flotation efficiency of the slag selection system is improved, the copper content of the tailings is reduced, and the direct recovery rate and the recovery rate of the system are improved; however, insufficient stirring, uneven distribution of materials and insufficient penetration cause incomplete reaction, the processing period is prolonged, and the production capacity is limited.

[0006] At present, the prior art one, the prior art two and the prior art three have problems of low smelting efficiency, reduced metal recovery rate and unstable product quality. In order to solve the above problems, the present application provides a top blowing furnace self-adaptive multi-hole rotary supersonic injection control method and system. SUMMARY

[0007] The main purpose of the present application is to provide a top blowing furnace self-adaptive multi-hole rotary supersonic injection control method and system to solve the problems of low smelting efficiency, reduced metal recovery rate and unstable product quality in the prior art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: An adaptive multi-hole rotary supersonic jet blowing control method for a top-blown furnace, comprising the following steps: The PLC control system identifies faults and abnormalities based on real-time status signals, automatically adjusts control strategies for actuators such as electric regulating ball valves, miniature solenoid valve groups, and multi-axis servo motors, and performs control analysis on fault and abnormal data. A molten pool behavior prediction model is constructed based on historical anomaly data, and the state signal after control is collected. The state signal after control is transmitted to the molten pool behavior prediction model to predict whether the state signal after control is qualified. If it is not qualified, a new control strategy is formed.

[0009] To achieve the above objectives, the present invention also provides the following technical solution: An adaptive multi-hole rotary supersonic jet blowing control system for a top-blown furnace is provided, which is applied to the aforementioned adaptive multi-hole rotary supersonic jet blowing control method for a top-blown furnace. The adaptive multi-hole rotary supersonic jet blowing control system for a top-blown furnace includes: a servo motor, a top-blown furnace, a static pressure sensor, a flow sensor, an electric regulating ball valve, a multi-hole spray gun, a dynamic pressure sensor, a vertical infrared thermometer, a grounding wire, an electric actuator, a multi-axis coupling, an integral rotator, a first zone, a second zone, a third zone, a fourth zone, a straight-through pressure valve, a multi-axis servo motor, a high-precision contact bearing, a pneumatic shut-off valve, an overflow valve, and a multi-hole nozzle. The servo motor is connected to the PCL control system via a grounding wire; a multi-hole spray gun is installed vertically inside the top-blown furnace, and a servo motor is installed on the top of the multi-hole spray gun; multiple static pressure sensors, flow sensors, electric regulating ball valves, and dynamic pressure sensors are installed on the spray gun pipes of the multi-hole spray gun, and are connected to the PCL control system; a vertical infrared detector detects the temperature in the top-blown furnace. The dynamic pressure sensor is installed at the top of the spray gun pipe, transmitting and receiving analog signals; the static pressure sensor is installed on the side near the spray gun pipe, receiving and transmitting analog signals; the flow sensor is installed on the side of the static pressure sensor and connected to the PLC control system via a grounding wire; the electric regulating ball valve is installed at the end of the spray gun pipe on one side, and an electric actuator is installed above the electric regulating ball valve.

[0010] As a further improvement of the present invention, the multi-hole spray gun is provided with a multi-hole nozzle, and the multi-hole spray gun includes a multi-axis coupling, the multi-axis connector is connected to an integral rotator, and the integral rotator drives the whole to rotate; the multi-hole spray gun nozzle is provided with multiple spray holes, which are divided into four sections, namely the first section, the second section, the third section and the fourth section; each section is equipped with an independent flow regulation circuit control mechanism for adjusting the local spray state of the section; The PLC system transmits data to a straight-through pressure valve, which controls a multi-axis servo motor. The multi-axis servo motor then outputs data back to the PLC system via an encoder and transmits it to a high-precision contact bearing, which in turn transmits it to a multi-hole nozzle. The multi-axis servo motor controls the first, second, third, and fourth zones respectively via pneumatic shut-off valves. Each of the first, second, third, and fourth zones is equipped with an overflow valve.

[0011] The edge computer of this invention sends the optimization results to the PLC control system. Through a PID control loop, it performs rapid closed-loop adjustment of actuators such as the electric regulating ball valve, micro solenoid valve group, and rotary servo motor based on real-time sensor feedback, achieving overall or zoned adjustment of the multi-hole nozzle. It can maintain stable total flow, nozzle opening, and rotational angular velocity under short-term dynamic conditions, effectively responding to instantaneous disturbances and emergency situations, ensuring the stable and reliable operation of the injection system. The mixed airflow in the pipeline, after passing through the rotatable multi-hole nozzle, impacts the pulverized coal particles in multiple directions, achieving a more uniform spatial distribution in the molten pool area. Compared to the traditional fixed single-channel injection mode, it effectively enhances the mass transfer, heat transfer, and chemical reaction processes between pulverized coal particles and the melt, providing crucial fluid dynamics support for efficient combustion and smelting in the furnace. Multiple high-precision sensors are deployed near the injection system to monitor the feeding status and operating parameters in real time. By combining MPC and ML algorithms for predictive analysis and optimized control, the shortcomings of existing technologies in real-time adjustment and intelligent optimization are overcome, significantly improving the system's intelligence level. The cloud server, based on historical operating data and real-time operating condition data, utilizes artificial neural networks for modeling and training, outputting globally optimized injection modes and control strategies, which are then distributed to edge computers and PLC systems via the network. The system can continuously update the MPC predictive model and control parameters, achieving coordinated control of remote global optimization and rapid local execution, further improving system operating efficiency and control accuracy. Attached Figure Description

[0012] Figure 1 This is a schematic flowchart of one embodiment of the adaptive multi-hole rotary supersonic jet blowing control method for top-blown furnace of the present invention; Figure 2 This is a detailed schematic diagram of an embodiment of the adaptive multi-hole rotary supersonic jet blowing control method for top-blown furnace of the present invention; Figure 3 This is a flowchart illustrating the steps of an embodiment of the adaptive multi-hole rotary supersonic jet control method for top-blown furnaces of the present invention, which involves real-time acquisition of status signals such as feed flow rate, pressure, temperature, and jet status. Figure 4This is a schematic flowchart illustrating the steps of an embodiment of the adaptive multi-hole rotary supersonic jet injection control method for top-blown furnaces of the present invention for controlling and analyzing fault and abnormal data. Figure 5 This is a schematic diagram of the steps in an embodiment of the adaptive multi-hole rotary supersonic injection control method for top-blown furnace of the present invention, which constructs a predictive model of molten pool behavior based on historical abnormal data. Figure 6 This is a schematic diagram of the feeding system of an embodiment of the adaptive multi-hole rotary supersonic jet blowing control device for top-blown furnace of the present invention; Figure 7 This is a schematic diagram of the sensor installation position in one embodiment of the adaptive multi-hole rotary supersonic jet control device for top-blown furnace of the present invention; Figure 8 This is a schematic diagram of the spray gun structure of an embodiment of the adaptive multi-hole rotary supersonic jet blowing control device for top-blown furnace of the present invention; Figure 9 This is a schematic diagram of the partitioned arrangement structure of an embodiment of the adaptive multi-hole rotary supersonic jet blowing control device for top-blown furnace of the present invention; Figure 10 This is a schematic diagram of the control rotatable circuit of an embodiment of the adaptive multi-hole rotary supersonic jet blowing control device for top-blown furnace of the present invention. Figure 11 This is a flowchart illustrating the operation of an embodiment of the adaptive multi-hole rotary supersonic jet injection control device for top-blown furnace of the present invention. Figure 12 This is a schematic diagram of the structure of an embodiment of the electronic device of the present invention; Figure 13 This is a schematic diagram of the structure of one embodiment of the storage medium of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0014] The terms "first," "second," and "third" used in this invention are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this invention are only used to explain the relative positional relationships and movements between components in a specific orientation (as shown in the accompanying drawings). If the specific orientation changes, the directional indications also change accordingly. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. 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 not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0015] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0016] like Figure 1 As shown, this embodiment provides an example of an adaptive multi-hole rotary supersonic jet injection control method for a top-blown furnace. In this embodiment, the adaptive multi-hole rotary supersonic jet injection control method for a top-blown furnace specifically includes the following steps: Step S1: Install sensors such as flow sensors, pressure sensors, and vertical infrared thermometers in the feed pipeline and injection system; construct a system to collect real-time status signals such as feed flow rate, pressure, temperature, and injection status; transmit the real-time status signals to the PLC control system via analog signals; Step S2: The PLC control system identifies faults and abnormalities based on real-time status signals, and automatically adjusts the control strategy for actuators such as electric regulating ball valves, miniature solenoid valve groups, and multi-axis servo motors; it also performs control analysis on fault and abnormal data. Step S3: Construct a molten pool behavior prediction model based on historical anomaly data, and collect the state signal after control; transmit the state signal after control to the molten pool behavior prediction model to predict whether the state signal after control is qualified; if it is not qualified, a new control strategy is formed.

[0017] Preferably, in this embodiment, the edge computer sends the optimization results to the PLC control system. Through a PID control loop, it performs rapid closed-loop adjustment of actuators such as the electric regulating ball valve, micro solenoid valve group, and rotary servo motor based on real-time sensor feedback, achieving overall or zoned adjustment of the multi-hole nozzle. This allows for stable total flow, nozzle opening, and rotational angular velocity under short-term dynamic conditions, effectively responding to instantaneous disturbances and emergency situations, ensuring the stable and reliable operation of the injection system. The mixed airflow in the pipeline, after passing through the rotatable multi-hole nozzle, impacts the pulverized coal particles in multiple directions, achieving a more uniform spatial distribution in the molten pool area. Compared to the traditional fixed single-channel injection mode, this effectively enhances the mass transfer, heat transfer, and chemical reaction processes between pulverized coal particles and the melt, providing crucial fluid dynamics support for efficient combustion and smelting within the furnace. Multiple high-precision sensors (pressure, flow, temperature) are deployed near the injection system to monitor the feeding status and operating parameters in real time. By combining MPC and ML algorithms for predictive analysis and optimized control, the shortcomings of existing technologies in real-time adjustment and intelligent optimization are overcome, significantly improving the system's intelligence level. The end server, based on historical operating data and real-time operating condition data, utilizes artificial neural networks for modeling and training, outputting globally optimized injection modes and control strategies, which are then distributed to edge computers and PLC systems via the network. The system can continuously update the MPC predictive model and control parameters, achieving coordinated control of remote global optimization and rapid local execution, further improving system operating efficiency and control accuracy (see appendix for details). Figure 2 ).

[0018] Furthermore, such as Figure 3 As shown, the process of real-time acquisition of status signals such as feed flow rate, pressure, temperature, and injection status in step S1 specifically includes the following steps: Step S11: Various sensors, such as pressure, flow rate, and temperature sensors, are installed in various areas of the feed pipeline and injection system to continuously monitor the temperature, flow rate, and pressure parameters inside the furnace, and transmit the collected data to the PLC control system in real time. Step S12: The edge computer and PLC system interact through the fieldbus to obtain real-time furnace condition and injection status data, run the calculation model, and output the attribution analysis results of abnormal operation through machine learning inference. Step S13: Based on the attribution analysis results, generate dynamic optimization control targets and send control commands to the PLC control system in real time.

[0019] Preferably, this embodiment establishes a predictive model of the complex physicochemical processes within the furnace, combines system constraints and control objectives, and dynamically calculates the optimal control sequence over a future period, ultimately providing precise control commands for actuators such as piezoelectric high-speed valves and servo motors. The top-blown furnace feeding system has strict physical and safety constraints; for example, the piezoelectric valve flow rate cannot exceed the upper limit, the nozzle angle has mechanical limits, and the temperature cannot exceed a safety threshold. The MPC must solve for the optimal control quantity while meeting these constraints; otherwise, equipment damage or safety accidents may occur. Constraints are divided into two categories: control input constraints (control quantity cannot exceed the range) and output constraints (operating conditions cannot exceed safety thresholds). Control input constraints must meet preset conditions; output constraints... : Output variables' safety / process range; for example, local furnace temperature, channel flow rate (below this value is prone to blockage, and above this value exceeds the equipment's rated load); perform optimization to obtain the optimal control sequence; solve using a quadratic programming algorithm or a nonlinear programming algorithm; after solving, obtain the optimal control sequence for the next Nc time steps; The control input constraints include:

[0020] The physical limits of the control quantity (e.g., the range of the opening degree K of a piezoelectric valve is 0%~100%) The mechanical limit of the nozzle angle θ is 0°~60°. The limit of the rate of change of the control quantity; Optimization solution, including:

[0021] By describing physical laws and quantifying control requirements through predictive models and objective functions, the system can achieve safety assurance under constraints and real-time adjustment through rolling optimization.

[0022] Furthermore, such as Figure 4 As shown, the process of controlling and analyzing fault abnormal data in step S2 specifically includes the following steps: Step S21: The sensor monitors the pressure difference before and after the nozzle in real time. When the detected value exceeds the set threshold, it is determined to be an operational abnormality such as nozzle blockage or abnormal spraying. Through PID loop error monitoring, when the control deviation cannot converge within the set time, it is determined to be an actuator failure. Step S22: After receiving the above abnormal signal, the PLC control system immediately executes the safety strategy: closes the electric regulating ball valve to cut off the total flow, and at the same time closes the nozzle solenoid valve to stop spraying; Step S23: The edge computer automatically adjusts the control strategy based on the abnormal state, reduces the total throughput setpoint, and reallocates the nozzle opening and rotation speed in an attempt to restore stable operation; the cloud server stores and analyzes the abnormal data.

[0023] Preferably, in this embodiment, the sensor monitors the pressure difference across the nozzle in real time. When the detected value exceeds a set threshold, it is determined to be an operational anomaly such as nozzle blockage or abnormal injection. Through PID loop error monitoring, when the control deviation cannot converge within a set time, it is determined to be an actuator failure. Upon receiving the above-mentioned abnormal signal, the PLC control system immediately executes a safety strategy: closing the electric regulating ball valve to cut off the total throughput, and simultaneously closing the nozzle solenoid valve to stop injection, thereby ensuring furnace safety. At the same time, the edge computer automatically adjusts the control strategy according to the abnormal state, reducing the total throughput setpoint and reallocating the nozzle opening and rotation speed to attempt to restore stable operation. The cloud server stores and analyzes the abnormal data, and updates the machine learning model based on historical and real-time information to optimize future control strategies, thereby reducing the risk of similar anomalies recurring.

[0024] Furthermore, the process of reallocating the nozzle opening and rotation speed in step S23 specifically includes the following steps: Step S231: The edge computer receives real-time status signals from the PLC control system, including the pressure difference monitoring value before and after the nozzle and the PID loop error value; the real-time status signals are processed by feature extraction: the pressure difference monitoring value before and after the nozzle is converted into a pressure difference dynamic gradient, and the PID loop error value is converted into an error accumulation intensity; the pressure difference dynamic gradient and the error accumulation intensity are fused by a feature to generate an anomaly fusion index; the anomaly fusion index is used to quantify the severity and type characteristics of the abnormal state; Step S232: The abnormal fusion index is processed by the total flux response to obtain the total flux attenuation factor; the total flux attenuation factor is combined with the current total flux setpoint to form a new total flux setpoint; the new total flux setpoint reflects the adaptation to abnormal conditions and ensures that the total flux is reduced to a stable operating range. Step S233: Based on the new total flux setting, an initial opening distribution profile is generated through the flux-opening mapping relationship; the anomaly fusion index modulates the initial opening distribution profile to generate a nozzle opening distribution sequence; simultaneously, the new total flux setting and the nozzle opening distribution sequence are sent to the speed coupler, which generates a rotational speed reference value that matches the current fluid flux and distribution pattern; the nozzle opening distribution sequence is used to configure the opening state of the micro solenoid valve group, and the rotational speed reference value is used to set the speed command of the multi-axis servo motor.

[0025] Preferably, in this embodiment, the pressure difference before and after the nozzle and the PID loop error are converted into gradient and cumulative intensity respectively and then fused to obtain an index that can objectively measure the degree and type of anomaly, enabling rapid capture and quantitative description of abnormal system states. The total flux attenuation factor generated based on the anomaly fusion index is synthesized with the original setpoint to form a new total flux setpoint. This allows the system to automatically reduce the overall flow rate when an anomaly occurs, maintaining a safe and stable operating range and avoiding overflow or instability caused by the anomaly. Based on the new total flux setpoint, an initial opening profile is obtained through flux-opening mapping. This profile is then modulated using the anomaly fusion index to obtain a nozzle opening distribution sequence that matches the current anomaly characteristics. The opening distribution sequence and the new total flux setpoint are fed into a speed coupler to generate a rotational speed reference value that matches the fluid flux and distribution pattern in real time, thereby driving a multi-axis servo motor to achieve precise speed control.

[0026] In summary, this embodiment, through adaptive adjustment of total flux and velocity driven by anomalies, enables the system to maintain stable fluid transport operation under various working conditions, reducing oscillations or shocks caused by abnormal fluctuations. Dynamic attenuation of total flux avoids unnecessary high-flow-rate operation, and, combined with optimal nozzle opening distribution, minimizes fluid dynamic resistance, thereby achieving higher energy efficiency. Real-time calculation of anomaly fusion indicators and rapid feature fusion and total flux synthesis allow the system to update parameters instantly upon anomaly occurrence, shortening the time link from detection to adjustment. The coupled generation mechanism of nozzle opening and rotational speed ensures their synchronous matching in space and time, improving the uniformity of fluid distribution and the consistency of processing / jetting quality. This embodiment, through real-time anomaly quantification, adaptive attenuation of total flux, and coordinated adjustment of nozzle opening and rotational speed, achieves comprehensive technical benefits such as high reliability, low energy consumption, rapid response, and high-precision control.

[0027] Furthermore, the process of generating the nozzle opening distribution sequence in step S233 specifically includes the following steps: Step S2331: Based on the new total flux setting value, select a corresponding reference aperture distribution pattern from the preset reference aperture spectrum; the reference aperture distribution pattern serves as the flux profile base, defining the basic aperture ratio of each nozzle under a given total flux. Step S2332: Based on the anomaly type and severity represented by the anomaly fusion index, output a set of corresponding opening adjustment coefficient sequences; the opening adjustment coefficient sequence constitutes an anomaly modulation vector, the value of which reflects the direction and magnitude of compensation or adjustment of the opening of each nozzle for the current anomaly. Step S2333: The basic opening ratio of each flux profile base is combined with the corresponding opening adjustment coefficient in the abnormal modulation vector to generate a set of nozzle opening instructions; the set of nozzle opening instructions is the nozzle opening distribution sequence, which is used to directly configure the opening state of the micro solenoid valve group.

[0028] Preferably, this embodiment achieves dynamic, precise, and anomaly-aware allocation of nozzle opening by selecting a reference opening spectrum, generating anomaly modulation vectors, and performing a synthesis operation of the two. This enables the system to perform targeted compensation for the opening of each nozzle based on real-time anomaly characteristics while maintaining the overall flux setting, ensuring the balance and stability of fluid distribution. At the same time, by directly generating opening commands that can be used for micro solenoid valves, the control link is simplified, the response speed and adjustment accuracy are improved, and the overall reliability, energy efficiency, and processing quality of the system are enhanced.

[0029] Furthermore, the process of generating a set of nozzle opening commands in step S2333 specifically includes the following steps: Step S23331: Interpret each basic opening ratio according to the new total flux setting value, and output a series of opening reference values ​​that characterize the absolute opening quantity; the opening reference values ​​constitute the initial operating reference for each nozzle under the current total fluid flux. Step S23332: Convert the aperture adjustment coefficient in the abnormal modulation vector into a modulation field with directional and intensity characteristics; the modulation field acts on each aperture reference value to generate a set of aperture correction values ​​containing abnormal compensation information; Step S23333: Perform saturation constraint processing on each opening correction value to ensure that its value falls within the effective operating range of the physical actuator; the value after saturation constraint processing is directly formatted as a nozzle opening command, and the set of nozzle opening commands constitutes the nozzle opening distribution sequence of the direct drive micro solenoid valve group.

[0030] Preferably, this embodiment achieves real-time, accurate, and constrained calculation of the nozzle opening command by superimposing an absolute opening reference based on the total flux with an abnormal modulation field and applying a physical saturation constraint to the generated correction value. This ensures that while meeting the overall flux requirements, effective compensation can be provided for abnormal states; the saturation constraint guarantees that the command always falls within the safe operating range of the actuator, preventing excessive or insufficient opening from causing system instability or hardware damage. Overall, this improves the reliability of the control command, the response speed, and the system's stable operation under abnormal conditions.

[0031] Furthermore, the process of converting the aperture adjustment coefficient in the anomalous modulation vector into a modulation field with directional and intensity characteristics in step S23332 specifically includes the following steps: Step S233321: Based on the sequence length of the abnormal modulation vector and the physical layout of the nozzle, generate a modulation field structure with corresponding dimensions and topological relationships; the modulation field structure provides a spatial framework for an empty, unassigned field. Step S233322: Read the value of each opening adjustment coefficient in the abnormal modulation vector. The sign of the value is interpreted as a direction command, which is used to mark the adjustment polarity in the modulation field structure. The absolute value of the value is interpreted as an intensity command, which is used to fill the adjustment energy level in the modulation field structure. The combination of the direction command and the intensity command forms the field strength and direction assignment in the modulation field structure. Step S233323: Based on the interaction logic between the nozzles, the field strength and direction are interpolated and diffused to finally output a spatially continuous and smooth modulation field; the modulation field acts on each aperture reference value to generate an aperture correction value.

[0032] Preferably, this embodiment maps the abnormal modulation vector into a spatial field structure that matches the nozzle layout, assigns direction and intensity to the field according to the coefficient sign and amplitude, and then uses the interaction rules between nozzles to interpolate and diffuse the field to form a continuous and smooth modulation field. It can uniformly express the directionality and intensity distribution of abnormal compensation in space, so that the correction of each opening reference value not only conforms to the local abnormality characteristics, but also maintains the overall smoothness, thereby improving the accuracy of nozzle opening adjustment, response consistency and system stability.

[0033] Furthermore, the process of interpolating and diffusing the field strength and direction assignments in step S233323 specifically includes the following steps: Step S2333231: Based on the physical layout of the nozzles, obtain the hydrodynamic coupling coefficient between each nozzle position and its adjacent nozzles, and generate a neighborhood coupling weight matrix; the neighborhood coupling weight matrix defines the strength and range of the mutual influence between each nozzle. Step S2333232: According to the propagation path defined by the neighborhood coupling weight matrix, the direction command and intensity command of each position are weighted and transmitted to its neighborhood to form the initial diffusion field; the initial diffusion field completes the initial spatial allocation of the discrete field values; Step S2333233: Apply spatial continuity constraints to the initial diffusion field, eliminate abrupt changes in the field quantity along the propagation path through iterative smoothing, and fuse the discrete field quantity distribution into a spatially continuous and smooth modulation field; the modulation field acts on each aperture reference value to generate an aperture correction value.

[0034] Preferably, in this embodiment Furthermore, such as Figure 5As shown, the process of constructing a melt pool behavior prediction model based on historical anomaly data in step S3 specifically includes the following steps: Step S31: Output the optimized future control strategy to the PLC control system, run the PID control loop, and perform rapid closed-loop adjustment of actuators such as electric regulating ball valve, micro solenoid valve group and rotary servo motor according to real-time sensor feedback, and make overall or zoned adjustments to the multi-hole nozzle. Step S32: The rotatable multi-hole nozzle is controlled by a multi-axis servo motor drive mechanism, which can change the injection direction in real time according to different working conditions in the furnace; enabling the supersonic airflow to impact the mixed gas transported by the pipeline at multiple angles and directions; the electric actuator in the electric regulating ball valve receives the model signal from the PLC system to realize the continuous adjustment of the valve opening. Step S33: Construct a molten pool behavior prediction model based on historical abnormal data, train the molten pool behavior prediction model based on historical operating data and real-time operating condition data; output the global optimized spraying mode and control strategy, and send them to the edge computer and PLC system to continuously update the MPC prediction model and control parameters.

[0035] Preferably, this embodiment operates a PID control loop, which performs rapid closed-loop regulation of actuators such as the electric regulating ball valve, miniature solenoid valve group, and rotary servo motor based on real-time sensor feedback. This allows for overall or zoned adjustment of the multi-orifice nozzle, ensuring that the total flow rate, nozzle opening, and rotational angular velocity remain stable under short-term dynamic conditions, thus responding to instantaneous disturbances and emergency situations. The rotatable multi-orifice nozzle is controlled by a multi-axis servo motor drive mechanism, which can change the injection direction in real time according to different operating conditions within the furnace. This enables the supersonic airflow to impact the mixed gas transported in the pipeline from multiple angles and directions. Simultaneously, the electric actuator in the electric regulating ball valve receives model signals from the PLC system to achieve continuous adjustment of the valve opening, allowing the flow rate in the pipeline to be adjusted according to operating conditions.

[0036] Furthermore, the process of constructing a melt pool behavior prediction model based on historical anomaly data in step S33 specifically includes the following steps: Step S331: Extract spatial features from real-time monitoring data; the melt pool behavior prediction model network includes an input layer, a convolutional layer, and a pooling layer; wherein, the input layer receives a real-time data matrix from the data acquisition layer. Each row represents the time series of a sensor; the convolutional layer extracts local features through convolution operations; Among them, the extraction of local features is as follows:

[0037] Step S332: The pooling layer reduces the dimensionality of the feature matrix and retains the main features; the final output feature map is used as the input of the RNN; the output feature map is used for time series modeling to capture the dynamic fluctuation of the molten pool surface; the output surface fluctuation trend and potential adverse working conditions are then analyzed. Among them, capturing the dynamic fluctuations of the molten pool surface is as follows:

[0038] In the formula, Currently in a hidden state. , Represents the weight matrix. For activation functions; The output layer is used to predict the future trend of liquid level fluctuations, and the corresponding formula is as follows: ; Step S333: The output layer is used to predict the future trend value of liquid surface fluctuations; by adjusting the weights W of the network to minimize the error between the output and the target, and using the idea of ​​least squares, the loss function is defined as the mean square error of the output vector. The mean square error of the output vector is:

[0039] Where represents the difference between the i-th element of the output vector and the i-th element of the label vector. This is specifically to make subsequent differentiation calculations more convenient.

[0040] Preferably, the core module for predicting operating conditions and generating optimal strategies in this embodiment works by establishing a predictive model of complex physicochemical processes within the furnace, combining system constraints and control objectives, dynamically calculating the optimal control sequence for a future period, and ultimately providing precise control commands to actuators such as piezoelectric high-speed valves and servo motors. The top-blown furnace feeding system has strict physical and safety constraints; the MPC must solve for the optimal control quantity while satisfying these constraints, otherwise it may lead to equipment damage or safety accidents. The constraints are divided into two categories: control input constraints and output constraints.

[0041] This embodiment also provides an embodiment of a top-blown furnace adaptive multi-hole rotary supersonic jet blowing control device. In this embodiment, the top-blown furnace adaptive multi-hole rotary supersonic jet blowing control device is applied to the top-blown furnace adaptive multi-hole rotary supersonic jet blowing control method as described in the above embodiment. The top-blown furnace adaptive multi-hole rotary supersonic jet blowing control device includes a servo motor 1, a top-blown furnace 2, a static pressure sensor 3, a flow sensor 4, an electric regulating ball valve 5, a multi-hole spray gun 6, a dynamic pressure sensor 7, a vertical infrared thermometer 8, a grounding wire 9, an electric actuator 10, a multi-axis coupling 11, an integral rotator 12, a first partition 13, a second partition 14, a third partition 15, a fourth partition 16, a straight-through pressure valve 17, a multi-axis servo motor 18, a high-precision contact bearing 19, a pneumatic shut-off valve 20, an overflow valve 21, and a multi-hole nozzle 22. Among them, such as Figure 6As shown, the servo motor 1 is connected to the PCL control system via the grounding wire 9; a multi-hole spray gun 6 is installed vertically inside the top-blown furnace 2, and the servo motor 1 is installed on the top of the multi-hole spray gun 6; multiple static pressure sensors 3, flow sensors 4, electric regulating ball valves 5 and dynamic pressure sensors 7 are installed on the spray gun pipe of the multi-hole spray gun, and are connected to the PCL control system; a vertical infrared detector detects the temperature in the top-blown furnace 2. like Figure 7 As shown, the dynamic pressure sensor 7 is installed at the top of the spray gun pipe, and emits and receives analog signals; the static pressure sensor 3 is installed on the side near the spray gun pipe, and receives and emits analog signals; the flow sensor 4 is installed on the side of the static pressure sensor 3, and is connected to the PLC control system via the grounding wire 9; the electric regulating ball valve 5 is installed at the end of the spray gun pipe, and an electric actuator 10 is installed above the electric regulating ball valve 5. like Figure 8 As shown, the multi-hole spray gun 6 is equipped with a multi-hole nozzle 22. The multi-hole spray gun 6 includes a multi-axis coupling 11, which is connected to an integral rotator 12. The integral rotator 12 drives the entire assembly to rotate. Figure 9 As shown, the nozzle of the multi-hole spray gun 6 is provided with multiple spray holes, which are divided into four sections: the first section 13, the second section 14, the third section 15 and the fourth section 16. Each section is equipped with an independent flow regulation circuit control mechanism to adjust the local spray state of the section. like Figure 10 As shown, the PLC system transmits data to the straight-through pressure valve 17 via AO, and the pressure valve controls the multi-axis servo motor 18; the multi-axis servo motor 18 outputs data to the PLC system via encoder feedback, and transmits it to the high-precision contact bearing 19, which in turn transmits it to the multi-hole nozzle 22; the multi-axis servo motor 18 controls the first partition 13, the second partition 14, the third partition 15, and the fourth partition 16 respectively via pneumatic shut-off valve 20; the first partition 13, the second partition 14, the third partition 15, and the fourth partition 16 are each equipped with an overflow valve 21.

[0042] Preferably, sensors such as flow sensors, pressure sensors, and vertical infrared thermometers are installed in the feed pipeline and injection system to collect information such as feed flow rate, pressure, temperature, and injection status in real time. All signals are transmitted to the PLC control system via analog signals and simultaneously uploaded to the edge computer and cloud server. The PLC control system, as the core control unit, runs a PID control loop and performs rapid closed-loop adjustment of actuators such as electric regulating ball valves, miniature solenoid valve groups, and multi-axis servo motors based on real-time sensor feedback. It promptly identifies various faults and abnormal conditions, ensuring that the total flow rate, nozzle opening, and rotational angular velocity remain stable under short-term dynamic conditions. This layer of control has high real-time performance to cope with instantaneous disturbances and emergency conditions. An edge computer is set up to run Model Predictive Control (MPC) and Machine Learning (ML) inference algorithms. The edge computer and the PLC control system interact via a fieldbus to transmit real-time collected operating condition information. After the edge computer performs calculations, it sends the optimized target setpoints and control strategies to the PLC control system. The PLC control system then corrects the parameters of the actuators, such as the electric regulating ball valve, the micro solenoid valve group, and the multi-axis servo motor, thereby achieving a high-speed, uniform, and adaptive feeding process. The cloud server has the function of running machine learning training and large-scale data optimization. Based on historical data and real-time operating conditions, the cloud performs in-depth modeling and optimization calculations to obtain the optimal injection mode, including the total feed throughput setting, the nozzle opening distribution scheme, and the nozzle rotation angular velocity. The optimization results are transmitted to the edge computer and the PLC control system via Ethernet, realizing multi-level intelligent control that combines remote collaborative optimization with local rapid execution.

[0043] Basic Principle of PID Control Loop: PID control loop refers to a closed-loop control strategy widely used in industrial automatic control systems. PID is an abbreviation for Proportional-Integral-Derivative (PID) control. PID control adjusts the actuator by continuously calculating the control deviation (the difference between the setpoint and the actual value) to make the system output follow the target value as closely as possible. Its three components are as follows: P (proportional): The control quantity is proportional to the deviation. Its main function is to respond quickly to deviations, but it is prone to generating steady-state errors.

[0044] I (Integral): The control quantity is proportional to the integral of the deviation and is used to eliminate long-term steady-state error.

[0045] D (derivative): The control quantity is proportional to the rate of change of the deviation. It is used to predict the deviation trend, improve the system response speed, and suppress oscillations.

[0046] The PID output formula is:

[0047] in: It is a control quantity (sent to the actuator). It is the deviation (set value - measured value). These are proportional, integral, and differential gains, respectively.

[0048] In summary, this embodiment utilizes various sensors, including pressure, flow rate, and temperature sensors, installed in different areas of the feed pipeline and injection system to continuously monitor the furnace's temperature, flow rate, and pressure parameters. The collected data is transmitted to the PLC control system in real time. Simultaneously, an edge computer interacts with the PLC system via a fieldbus to acquire real-time furnace and injection status data, run a computational model, and use machine learning inference to analyze the causes of abnormal operations. This analysis, combined with the MPC algorithm, generates dynamic optimization control objectives, and precisely sends control commands to the PLC control system in real time, achieving a "data-driven, predictive inference, and optimization decision-making" approach. The intelligent control closed loop ensures the dynamic stability and parameter controllability of the smelting process, and outputs optimized control targets in a timely manner. Sensors monitor the pressure difference before and after the nozzle in real time. When the detected value exceeds the set threshold, it is determined to be an operational abnormality such as nozzle blockage or abnormal injection. Through PID loop error monitoring, when the control deviation cannot converge within a set time, it is determined to be an actuator failure. After receiving the above abnormal signals, the PLC control system immediately executes the safety strategy: closes the electric regulating ball valve to cut off the total throughput, and closes the nozzle solenoid valve to stop injection, thereby ensuring furnace safety. At the same time, the edge computer automatically adjusts the control strategy according to the abnormal state, reduces the total throughput set value, and redistributes the nozzle opening and rotation speed to try to restore stable operation. The cloud server stores and analyzes abnormal data, and updates the machine learning model based on historical and real-time information to optimize future control strategies and reduce the risk of similar anomalies recurring. The edge computer's optimization results are sent to the PLC control system, which runs a PID control loop. Based on real-time sensor feedback, it performs rapid closed-loop regulation of actuators such as the electric regulating ball valve, micro-solenoid valve group, and rotary servo motor, and adjusts the multi-orifice nozzle as a whole or in sections to ensure that the total flow, nozzle opening, and rotational angular velocity remain stable under short-term dynamic conditions, responding to instantaneous disturbances and emergency situations. The rotatable multi-orifice nozzle is controlled by a multi-axis servo motor drive mechanism, which can change the injection direction in real time according to different operating conditions in the furnace. This allows the supersonic airflow to impact the mixed gas transported in the pipeline from multiple angles and directions. Simultaneously, the electric actuator in the electric regulating ball valve receives model signals from the PLC system to achieve continuous adjustment of the valve opening, allowing the flow rate in the pipeline to adjust according to operating conditions. Based on historical operating data and real-time operating condition data, the cloud server uses artificial neural network methods for modeling and training, outputting a globally optimized injection mode and control strategy. By distributing data over the network to edge computers and PLC systems, the MPC prediction model and control parameters are continuously updated, enabling remote global optimization and rapid local execution in tandem (see appendix for details). Figure 11 ).

[0049] The adaptive multi-hole rotary supersonic jet blowing system for top-blown furnaces provided in this embodiment mainly includes the following functional units: 1) Rotatable multi-hole supersonic jet unit: Rotatable multi-hole spray gun mechanism: A rotatable multi-hole supersonic nozzle is installed at the end of the nozzle pipe. The overall rotation of the nozzle is driven by a multi-axis servo motor to achieve controllable angular velocity rotation. The nozzle has multiple spray holes, divided into four zones: I, II, III, and IV. Each zone is equipped with an independent flow regulation circuit control mechanism to adjust the local spray state of the zone; Servo motor drive mechanism: A multi-axis servo motor is used, equipped with a high-precision angle encoder (resolution ≤ 0.1°), to drive the nozzle to adjust the overall angle and zone angle within the range of −15° to +15°, covering different spray areas in the furnace; Electric regulating ball valve: The electric actuator is driven by a motor, combined with a position feedback device, to achieve continuous adjustment of the valve opening; 2) Intelligent monitoring and control unit: The multi-level intelligent control architecture, which combines sensor perception, PLC system control (Programmable Logic Controller), edge computing (MPC (Model Predictive Control) combined with ML machine inference) and cloud learning, has the following specific process: Data Acquisition Layer: Sensors such as flow sensors, pressure sensors, and vertical infrared thermometers are installed in the feed pipeline and injection system to collect information such as feed flow rate, pressure, temperature, and injection status in real time. All signals are transmitted to the PLC control system via analog signals and simultaneously uploaded to the edge computer and cloud server. PLC Control Layer: The PLC control system serves as the core control unit, running a PID control loop. Based on real-time sensor feedback, it performs rapid closed-loop regulation of actuators such as electric regulating ball valves, miniature solenoid valve assemblies, and multi-axis servo motors. It promptly identifies various faults and abnormal conditions, ensuring that total flow, nozzle opening, and rotational angular velocity remain stable under short-term dynamic conditions. This layer of control possesses high real-time performance, enabling it to handle transient disturbances and emergency situations.

[0050] Edge Optimization Layer: An edge computer is set up to run Model Predictive Control (MPC) and Machine Learning (ML) inference algorithms. The edge computer interacts with the PLC control system via a fieldbus, transmitting real-time acquired operating condition information. After calculation and execution, the edge computer sends the optimized target setpoints and control strategies to the PLC control system. The PLC control system then corrects the parameters of the actuators, such as the electric regulating ball valve, the miniature solenoid valve assembly, and the multi-axis servo motor, thereby achieving a high-speed, uniform, and adaptive feeding process.

[0051] Cloud-based learning layer: The cloud server has the capability to run machine learning training and large-scale data optimization. Based on historical data and real-time operating conditions, the cloud performs in-depth modeling and optimization calculations to obtain the optimal injection mode, including the total feed throughput setting, nozzle opening allocation scheme, and nozzle rotational angular velocity. The optimization results are transmitted via Ethernet to the edge computer and PLC control system, achieving multi-level intelligent control that combines remote collaborative optimization with rapid local execution.

[0052] like Figure 12 As shown, this embodiment provides an example of an electronic device 23, which includes a processor 231 and a memory 232 coupled to the processor 231. The memory 232 stores program instructions for implementing the adaptive multi-orifice rotary supersonic jet blowing control method for top-blown furnaces according to any of the above embodiments. The processor 231 executes the program instructions stored in the memory 232 to perform adaptive multi-orifice rotary supersonic jet blowing control for top-blown furnaces. The processor 231 can also be referred to as a CPU (Central Processing Unit). The processor 231 may be an integrated circuit chip with signal processing capabilities. The processor 231 can 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 devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor.

[0053] Furthermore, Figure 13 This is a schematic diagram of the structure of a storage medium according to an embodiment of this application. The storage medium 24 of this embodiment stores program instructions 241 capable of implementing all the above methods. These program instructions 241 can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0054] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, apparatuses, or units, and may be electrical, mechanical, or other forms. Additionally, the functional units in the various embodiments of this 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 integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this invention and do not limit the patent scope of this invention. Any equivalent structural or procedural transformations made using the content of this specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this invention.

[0055] The specific embodiments of the invention have been described in detail above, but these are merely examples, and the 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 this invention. Therefore, all equivalent transformations, modifications, and improvements made without departing from the spirit and principles of this invention should be included within the scope of this invention.

Claims

1. A method for adaptive multi-hole rotary supersonic injection control of a top- blown converter, characterized in that, The top-blown furnace adaptive multi-hole rotary supersonic injection control method comprises the following steps: The PLC control system identifies fault abnormal conditions according to real-time state signals, and automatically adjusts the control strategy of the electric regulating ball valve, the micro electromagnetic valve group and the actuator of the multi-axis servo motor; the fault abnormal data is controlled and analyzed; Based on the historical abnormal data, a molten pool behavior prediction model is constructed, and the controlled state signals are collected; the controlled state signals are transmitted to the molten pool behavior prediction model for prediction whether the controlled state signals are qualified; if not, a new control strategy is formed.

2. The top-submerged lance self-adaptive multi-hole rotary ultrasonic injection control method according to claim 1, characterized in that, The process of controlling and analyzing the fault abnormal data comprises the following steps: The sensor monitors the pressure difference before and after the nozzle in real time, and when the detected value exceeds the set threshold, it is determined that the nozzle is blocked or abnormally operated; through PID loop error monitoring, when the control deviation cannot converge within the set time, it is determined that the actuator is faulty; After receiving the above abnormal signals, the PLC control system executes a safety strategy: closes the electric regulating ball valve to cut off the total flux, and closes the injection hole electromagnetic valve to stop injection; The edge computer automatically adjusts the control strategy according to the abnormal state, reduces the total flux set value, and reallocates the injection hole opening degree and the rotation speed to try to restore stable operation; the cloud server stores and analyzes the abnormal data.

3. The top-submerged lance self-adaptive multi-hole rotary ultrasonic injection control method according to claim 2, characterized in that, The process of reallocating the injection hole opening degree and the rotation speed comprises the following steps: The edge computer receives real-time state signals from the PLC control system, including nozzle pressure difference monitoring values and PID loop error values; the real-time state signals are processed through feature extraction: the nozzle pressure difference monitoring values are converted into pressure difference dynamic gradients, and the PID loop error values are converted into error accumulation intensities; the pressure difference dynamic gradient and the error accumulation intensity are fused to generate an abnormal fusion index; the abnormal fusion index is used to quantify the severity and type characteristics of the abnormal state; The abnormal fusion index is processed through total flux response to obtain a total flux attenuation factor; the total flux attenuation factor is combined with the current total flux set value to form a new total flux set value; According to the new total flux set value, an initial opening degree distribution profile is generated through the flux-opening degree mapping relationship; the abnormal fusion index modulates the initial opening degree distribution profile to generate an injection hole opening degree distribution sequence; at the same time, the new total flux set value and the injection hole opening degree distribution sequence are sent into a rotation speed coupler to generate a rotation speed reference value matched with the current fluid flux and distribution form; the injection hole opening degree distribution sequence is used to configure the opening degree state of the micro electromagnetic valve group, and the rotation speed reference value is used to set the rotation speed instruction of the multi-axis servo motor.

4. The top-submerged lance adaptive multi-hole rotary ultrasonic injection control method according to claim 3, characterized in that, The process of generating the injection hole opening degree distribution sequence comprises the following steps: According to the new total flux set value, a corresponding reference opening degree distribution mode is selected from the preset reference opening degree spectrum; the reference opening degree distribution mode is used as a flux profile base to define the basic opening degree proportion of each injection hole under a given total flux; According to the type and severity of the anomaly represented by the anomaly fusion index, a set of corresponding opening adjustment coefficient sequences is output; the opening adjustment coefficient sequences form an anomaly modulation vector, and the values thereof reflect the direction and amplitude of the compensation or adjustment of each injection orifice opening for the current anomaly; Each base opening ratio of the flux profile base is subjected to a synthetic operation with the corresponding opening adjustment coefficient in the anomaly modulation vector to generate a set of injection orifice opening instruction sets; the injection orifice opening instruction sets are the injection orifice opening distribution sequences for directly configuring the opening state of the micro electromagnetic valve group.

5. The top-submerged lance adaptive multi-hole rotary ultrasonic injection control method according to claim 4, characterized in that, The process of generating a set of injection orifice opening instruction sets includes the following steps: Each base opening ratio is interpreted according to the new total flux set value to output a series of opening reference values representing absolute opening quantities; the opening reference values form the initial operation reference of each injection orifice under the current total fluid flux; The opening adjustment coefficients in the anomaly modulation vector are converted into a modulation action field with directional and intensity characteristics; the modulation action field acts on each opening reference value to generate a set of opening correction values containing anomaly compensation information; Each opening correction value is subjected to saturation constraint processing to ensure that its value falls within the effective operation interval of the physical actuator; the value after saturation constraint processing is directly formatted as an injection orifice opening instruction, and the set of orifice opening instructions constitutes the injection orifice opening distribution sequence of the directly driven micro electromagnetic valve group.

6. The top-submerged lance adaptive multi-hole rotary ultrasonic injection control method according to claim 5, characterized in that, The process of converting the opening adjustment coefficients in the anomaly modulation vector into a modulation action field with directional and intensity characteristics includes the following steps: According to the sequence length of the anomaly modulation vector and the physical layout of the injection orifice, a modulation field domain structure with corresponding dimensions and topological relationships is generated; The values of each opening adjustment coefficient in the anomaly modulation vector are read; the sign of the value is interpreted as a direction instruction for marking the adjustment polarity in the modulation field domain structure; the absolute value of the value is interpreted as an intensity instruction for filling the adjustment energy level in the modulation field domain structure; the combination of the direction instruction and the intensity instruction forms the assignment of field strength and direction in the modulation field domain structure; According to the interaction logic between the injection orifices, the field strength and direction assignment is interpolated and diffused to finally output a modulation action field that is continuous and smooth in space; the modulation action field acts on each opening reference value to generate an opening correction value.

7. The top-submerged lance adaptive multi-hole rotary ultrasonic injection control method according to claim 6, characterized in that, The process of interpolating and diffusing the field strength and direction assignment includes the following steps: According to the physical layout of the injection orifice, the fluid dynamic coupling coefficients between each injection orifice position and its adjacent injection orifice are obtained to generate a neighborhood coupling weight matrix; the neighborhood coupling weight matrix defines the strength and range of the mutual influence between the injection orifices; According to the propagation path defined by the neighborhood coupling weight matrix, the direction instruction and the intensity instruction of each position are weighted and transmitted to its neighborhood to form an initial diffusion field; the initial diffusion field completes the initial spatial distribution of discrete field values; The spatial continuity constraint is applied to the initial diffusion field, and the iteration smoothing process is used to eliminate the mutation of the field quantity on the propagation path, and the discrete field quantity distribution is fused into a modulation action field that is continuous and smooth in space; the modulation action field acts on each opening reference value to generate an opening correction value.

8. The top-submerged lance adaptive multi-hole rotary ultrasonic injection control method according to claim 1, characterized in that, The process of constructing a molten pool behavior prediction model based on historical abnormal data includes the following steps: The optimized future control strategy is output to the PLC control system, the PID control loop is run, and the electric regulating ball valve, the micro electromagnetic valve group, and the rotating servo motor are quickly closed-loop regulated according to real-time sensor feedback; the multi-hole nozzle is adjusted as a whole or in zones; The multi-hole nozzle is controlled by the multi-axis servo motor driving mechanism, and the jet direction is changed in real time according to different working conditions in the furnace; the supersonic airflow can impact the mixed gas transported by the pipeline from multiple angles; the electric actuator of the electric regulating ball valve receives the model signal of the PLC system to realize continuous adjustment of the valve opening; The molten pool behavior prediction model is constructed based on historical abnormal data, and the molten pool behavior prediction model is trained based on historical operation data and real-time working condition data; the global optimization jetting mode and control strategy are output to the edge computer and the PLC system, and the MPC prediction model and control parameters are continuously updated.

9. A control system for adaptive multi-hole rotary supersonic injection of a top- blown vessel, applied to the control method for adaptive multi-hole rotary supersonic injection of a top-blown vessel according to any one of claims 1 to 8, characterized in that, The top-blown furnace adaptive multi-hole rotary supersonic jetting control system comprises a servo motor, a top-blown furnace, a static pressure sensor, a flow sensor, an electric regulating ball valve, a multi-hole lance, a dynamic pressure sensor, a vertical infrared temperature detector, a grounding wire, an electric actuator, a multi-axis coupling, an overall rotator, a first zone, a second zone, a third zone, a fourth zone, a straight-through pressure valve, a multi-axis servo motor, a high-precision contact bearing, a pneumatic stop valve, an overflow valve, and a multi-hole nozzle. The servo motor is connected to the PCL control system through the grounding wire; the multi-hole lance is vertically arranged in the top-blown furnace, and the servo motor is arranged on the top of the multi-hole lance; a plurality of static pressure sensors, flow sensors, electric regulating ball valves, and dynamic pressure sensors are arranged on the lance pipe of the multi-hole lance and connected to the PCL control system; the vertical infrared detector detects the temperature in the top-blown furnace; The dynamic pressure sensor is installed on the top of the lance pipe and emits an analog signal for analog signal reception; the static pressure sensor is installed on one side close to the lance pipe and emits an analog signal for analog signal reception; the flow sensor is installed on one side of the static pressure sensor and connected to the PLC control system through the grounding wire; the electric regulating ball valve is installed at the end of one side of the lance pipe, and the electric actuator is installed above the electric regulating ball valve.

10. The top-submerged lance self-adaptive multi-hole rotary ultrasonic injection control system according to claim 9, characterized in that, The multi-hole lance comprises a multi-axis coupling, a multi-axis connector, an overall rotator, and a multi-hole nozzle; the multi-hole nozzle comprises a plurality of jet holes and is divided into four zones, namely a first zone, a second zone, a third zone, and a fourth zone; each zone is equipped with an independent flow regulation circuit control mechanism for adjusting the local jetting state of the zone. The PLC system transmits data to the straight-through pressure valve, the pressure valve controls the multi-axis servo motor, the multi-axis servo motor feeds back the data to the PLC system through the encoder and transmits the data to the high-precision contact bearing, the high-precision contact bearing transmits the data to the multi-hole nozzle, the multi-axis servo motor controls the first sub-area, the second sub-area, the third sub-area and the fourth sub-area through the pneumatic stop valve respectively, and the first sub-area, the second sub-area, the third sub-area and the fourth sub-area are respectively provided with overflow valves.

Images