Steel production equipment self-adaptive regulation and control method and system based on instantaneous energy efficiency characteristic feedback

Abstract

The application relates to the field of steel metallurgy automatic control, and discloses a steel production equipment self-adaptive regulation and control method and system based on instantaneous energy efficiency characteristic feedback. The method synchronously collects equipment operation states and material flow data in a millisecond level period, and solves an instantaneous characteristic value representing unit effective output energy consumption intensity by using an energy consumption-material coupling model; under the limitation of a process stability constraint model, based on the comparison result of the instantaneous characteristic value and a dynamic reference threshold value, a physical correction instruction is automatically generated to drive an actuator to execute closed-loop regulation. The application converts the energy efficiency index into a real-time process control variable, realizes dynamic balance of energy efficiency optimization and process stability, and deals with irreversible material loss through cross-process compensation.

Description

Technical Field

[0001] This application relates to the field of steel production process control technology, and in particular to an adaptive control method and system for steel production equipment based on instantaneous energy efficiency characteristic feedback. Background Technology

[0002] As a typical process industry, the steel industry's hot rolling production line encompasses multiple continuous processes, including heating, roughing, finishing, laminar cooling, and coiling. These processes are closely coupled in terms of energy and material flow. In actual production, hot rolling production lines face various typical scenarios requiring energy efficiency control.

[0003] The first typical scenario is the problem of high energy consumption due to empty combustion in the heating furnace area. When the material flow slows down due to equipment maintenance, roll changing, or billet quality inspection in the previous process, the billet stays in the furnace for a longer period of time, but the gas regulating valve and combustion fan maintain their original high-level supply. At this time, although the billet in the furnace has reached or exceeded the target heating temperature, the continuous high-level combustion causes a large amount of heat energy to be lost in the form of flue gas waste heat, resulting in significant ineffective energy consumption.

[0004] The second typical scenario is the increased unit energy consumption caused by material volume loss. In the roughing process, if the amount of head and tail trimming exceeds expectations or the degree of oxidation and burn-off is greater than the design value, the actual volume of the intermediate billet will be smaller than the theoretical value. Since the preceding heating process has already consumed the corresponding energy, and the effective material output is reduced, the unit energy consumption of this batch of products will inevitably increase. This kind of material loss that has already occurred is an irreversible process, and traditional control methods are powerless to address it.

[0005] The third typical scenario is the problem of uneven load distribution in the rolling mill area. During the operation of the continuous rolling mill, due to the complex disturbances of the rolling process, the main motors of some stands may occasionally deviate from the high-efficiency operating range, exhibiting an inefficient and high-consumption state with high-frequency current fluctuations, resulting in higher overall power consumption of the unit.

[0006] The fourth typical scenario involves the discrepancy between carbon inventory data and actual emissions. Carbon emissions in steel production primarily originate from direct emissions from fuel combustion and indirect emissions from purchased electricity. These emissions have a deterministic physical relationship with fuel and electricity consumption based on stoichiometry and the grid carbon emission factor. However, in actual production, discrepancies arise between the actual and nominal calorific values ​​of fuel, the actual and designed operating efficiencies of equipment, and errors in the actual and statistical quantities of materials processed. This leads to significant deviations between carbon inventory data calculated based on fixed emission factors and statistical data and carbon emissions based on actual physical processes. Such discrepancies are difficult to detect and trace in a timely manner without real-time physical measurement and verification methods.

[0007] Regarding the aforementioned scenarios, existing technologies have the following limitations. First, while existing energy management systems can perform energy consumption statistics and analysis, their action lags behind the production process itself, typically using minute-level or even batch-level statistical cycles, and the output is limited to reports or alarm signals. They cannot directly participate in real-time equipment control, leading to energy waste being discovered only after it has already occurred. Second, while existing process control systems can achieve closed-loop control of physical quantities such as temperature, pressure, and thickness, their control objectives only target the physical quality indicators of the product, lacking a real-time feedback loop for "energy efficiency," and thus failing to identify suboptimal operating states such as "high quality but high energy consumption." Third, existing technologies lack an effective mechanism for dynamically balancing energy efficiency optimization and process stability. Control actions solely pursuing energy efficiency optimization may lead to process quality problems such as uneven furnace temperature, plate defects, or equipment overheating. Furthermore, existing control logic is usually limited to isolated optimization of a single process, lacking cross-process energy efficiency collaborative compensation capabilities. Furthermore, existing carbon accounting systems typically use fixed emission factors and batch-level statistical data for ex-post accounting, lacking physical measurement data based on the real-time operating status of production equipment as a verification basis. When there is a discrepancy between the carbon accounting inventory data and actual emissions, due to the lack of physical sensor data with millisecond-level time resolution, it is difficult to trace the specific source of the discrepancy, whether it is a fluctuation in fuel calorific value, equipment efficiency decline, or material metering error. Therefore, it is impossible to effectively achieve closed-loop control of carbon emissions by adjusting equipment operating parameters.

[0008] Therefore, there is an urgent need for a new technical solution that can sense energy efficiency deviations in real time on a millisecond timescale, automatically generate physical correction instructions under the constraint of ensuring process stability, and achieve coordinated optimization of energy efficiency throughout the entire process. At the same time, the technical solution should also be able to provide a verification basis with a clear physical source for carbon accounting inventory data based on real-time collected physical sensor data, and support the active control of carbon emissions by adjusting equipment operating parameters when deviations are identified. Summary of the Invention

[0009] The purpose of this application is to provide an adaptive control method and system for steel production equipment based on instantaneous energy efficiency characteristic feedback, so as to solve the problems mentioned in the background art.

[0010] This application discloses an adaptive control method for steel production equipment based on instantaneous energy efficiency characteristic feedback, comprising the following steps: S1: By deploying sensor arrays in key process sections of the steel production line, equipment operating status data is collected synchronously at millisecond intervals. and material flow status data ; The device operating status data It includes at least: the instantaneous flow rate of the fuel medium and the current value of the drive motor; The material flow status data At least including: temperature distribution and geometric dimension data of the work-in-process obtained through infrared imaging or laser scanning; S2: Utilize the edge computing unit to process the device operating status data collected in step S1. Material flow status data Input a preset energy consumption-material coupling model and calculate the instantaneous characteristic value representing the energy consumption intensity per unit of effective output within the current sampling period. When the material input remains constant but the effective output decreases, the instantaneous characteristic value... It exhibits a transient, pulse-like increase; S3: The instantaneous characteristic value obtained in step S2 is processed. Compare with dynamic benchmark thresholds to determine if there is any energy efficiency deviation; When the instantaneous feature value is detected When the threshold exceeds the upper limit of the dynamic baseline threshold, the process stability constraint model is further invoked to calculate the equipment safety index in real time based on the data collected in step S1. Verify whether the current device is within the allowable adjustment safety margin range; S4: When step S3 determines the instantaneous feature value Abnormal deviation and the equipment safety indicators When the safety margin requirements are met, physical correction instructions for the actuators of the production equipment are generated based on the control algorithm and sent to the process control system through the industrial communication interface; The physical correction command drives the actuator to adjust the physical operating parameters, so that the instantaneous characteristic value The dynamic baseline threshold range is regressed to form a closed-loop feedback.

[0011] In a preferred embodiment, in step S1: The device operating status data This also includes: the torque of the main motor of the rolling mill, the frequency of the combustion fan, the power of the cooling water pump, and the pressure of the descaling pump; The material flow status data It also includes: instantaneous throughput based on roller conveyor weighing sensors and material conveying speed based on speed encoders.

[0012] In a preferred embodiment, in step S2, the instantaneous feature value Solve it as follows: in: The power consumption of the equipment during the current sampling period is obtained by multiplying the instantaneous flow rate of the fuel medium by the preset or online detected calorific value of the fuel, and summing the power consumption of the drive motor. The effective material output rate within the current sampling period is calculated from the real-time cross-sectional area of ​​the work-in-process obtained by combining the roller conveyor weighing signal and speed measurement signal with laser diameter measurement. The instantaneous feature value The unit is kJ / kg, which represents the energy intensity per unit of effective output.

[0013] In a preferred embodiment, in step S3, the dynamic benchmark threshold is automatically generated based on historical stable operating condition data, and its generation method is as follows: Instantaneous characteristic values ​​of qualified operating conditions within the most recent preset time window Perform sliding window statistics to calculate the mean and standard deviation within the window; The upper limit of the threshold is set by adding a preset multiple of the standard deviation to the mean; When the equipment is detected to be in an unsteady state of switching steel grades or changing rolls, the update of the dynamic reference threshold is paused.

[0014] In a preferred embodiment, in step S3, the process stability constraint model includes the following constraint indices: Furnace temperature uniformity variance constraint: Calculate the temperature variance based on the temperature data of multiple thermocouples in the furnace, and determine whether the temperature variance is less than the upper limit of variance set in the process specification. Plate shape straightness constraint: Based on the real-time feedback signal of the plate shape instrument, the wave shape index is obtained, and it is determined whether the wave shape index is within the preset controlled range; Motor temperature rise rate constraint: The temperature rise rate is calculated based on data from the motor winding temperature sensor, and it is determined whether the temperature rise rate is lower than the safety threshold. The physical correction instruction will only be issued to the actuator if the generated physical correction instruction is predicted by simulation to not cause the above-mentioned constraint indicators to exceed the limit.

[0015] In a preferred embodiment, in step S4, the actuator includes at least one of the following: The gas regulating valve of the heating furnace is used to adjust the valve opening to control the fuel supply. The frequency converter of the combustion blower is used to adjust the blower frequency to control the combustion air volume; The roll gap servo mechanism of the rolling mill is used to adjust the roll gap setpoint to control the reduction amount; The frequency converter of the cooling system's water pump is used to adjust the pump's operating frequency to control the cooling water flow rate.

[0016] In a preferred embodiment, for the heating furnace area, the control strategy in step S4 includes: When step S1 detects that the material flow slows down due to a decrease in the speed of the furnace feed rollers, but the gas flow rate remains high, the instantaneous characteristic value calculated in step S2 is... The temperature rises, indicating a state of high energy consumption due to idling; Under the premise that the surface temperature of the steel billet in the furnace has reached the upper limit of the process and the center temperature meets the process requirements in step S3, it is determined that the stability constraint allows for cooling. The physical correction instructions include: reducing the opening of the gas regulating valve in the heat exchange section, and / or reducing the output frequency of the combustion fan inverter, thereby reducing excess air heat loss and reducing the supply of ineffective calorific value.

[0017] In a preferred embodiment, step S4 includes a cross-process energy consumption dynamic compensation strategy for material volume loss: When the laser scanning data in step S1 indicates that the measured volume of the intermediate billet at the roughing mill exit is less than the theoretically set threshold, it indicates that excessive cutting at the beginning and end or significant oxidation and burning loss has led to irreversible volume loss of the material, causing the instantaneous characteristic value of the unit effective output energy consumption intensity of this batch of products calculated in step S2 to be affected. Increase; To offset the increased energy consumption per unit caused by irreversible material losses in the preceding stages, the physical correction command is applied to the subsequent laminar flow cooling process: Under the condition of meeting the winding temperature process deviation constraint, the low-pressure section cooling manifold is automatically and preferentially turned on, and the inverter output frequency of the high-pressure water supply pump is reduced or turned off. By reducing the electrical energy consumption in the cooling process, material loss in the preceding processes can be physically compensated, thereby achieving energy efficiency optimization throughout the entire process.

[0018] In a preferred embodiment, the control strategy in step S4 for the rolling region includes: Based on the main motor current data of each rack collected in step S1, when an abnormal high-frequency fluctuation is detected in the main motor current of a rack, it is identified that the instantaneous energy efficiency ratio of the rack is lower than the preset lower limit threshold of energy efficiency ratio, and it is in the low-efficiency and high-consumption operation zone. Using an automatic thickness control system, while maintaining a constant total reduction rate, a portion of the reduction load of this frame is dynamically distributed to adjacent high-efficiency frames. This allows the entire unit to operate within the high-efficiency range of the motor, reducing overall power consumption.

[0019] In a preferred embodiment, a real-time verification step for the carbon accounting inventory data is also included: Based on the instantaneous flow rate of the fuel medium and the power of the drive motor collected in step S1, and combined with the fuel carbon emission factor and the power grid carbon emission factor, the instantaneous carbon emission intensity is calculated. The instantaneous carbon emission intensity is integrated over a preset statistical period to obtain the cumulative carbon emissions based on measured physical data. The cumulative carbon emissions are compared with the declared carbon emissions in the carbon accounting inventory. When the deviation rate exceeds a preset threshold, a verification anomaly identifier is generated, and the source of the deviation is traced based on the physical data collected in step S1.

[0020] In a preferred embodiment, the control algorithm is at least one of a model predictive control algorithm or a PID control algorithm; When using model predictive control algorithms: With the instantaneous characteristic value The optimization objective is to minimize the deviation from the dynamic baseline threshold. The constraints are furnace temperature uniformity variance constraint, plate flatness constraint, and motor temperature rise rate constraint in the process stability constraint model. The optimal control quantity is solved in the prediction time domain, and the deviation between the measured output and the predicted output is used to correct the model state online.

[0021] This application also discloses an adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback, including: The multi-source sensing module is used to connect to the field sensor group through the sensor interface to collect equipment operating status data in real time. and material flow status data The device operating status data This includes at least the instantaneous flow rate of the fuel medium and the current value of the drive motor; the material flow status data. At least include data on the temperature distribution, geometric dimensions, and instantaneous material throughput of the product; The feature calculation module, connected to the multi-source sensing module, is configured to run an energy consumption-material coupling model, mapping the data collected by the multi-source sensing module into instantaneous feature values. ; An optimized decision controller, connected to the feature calculation module, is configured to, under the constraints of a process stability constraint model, calculate the instantaneous feature values. The control variables are calculated based on the deviation of instantaneous characteristic values ​​by comparing with dynamic benchmark thresholds; the process stability constraint model includes furnace temperature uniformity variance constraint, plate flatness constraint and motor temperature rise rate constraint. The execution drive interface is connected to the optimization decision controller to convert the control variables into standard PLC control instructions and output them to the heating furnace valve positioner, motor frequency converter or hydraulic servo system to adjust the physical operating parameters of the equipment; the execution result is fed back to the feature calculation module through the multi-source sensing module to form a closed loop.

[0022] In a preferred embodiment, the feature solving module is deployed in an edge computing unit close to the production site; The multi-source sensing module and the feature calculation module transmit data using an industrial communication protocol. The execution drive interface communicates with the process control system via a PLC or DCS interface to issue commands.

[0023] The adaptive control method and system for steel production equipment based on instantaneous energy efficiency characteristic feedback provided in this application can achieve the following technical effects through the coordinated operation of four steps: real-time sensing of multi-source physical data, instantaneous energy efficiency characteristic calculation, dual-constraint optimization decision-making, and closed-loop feedback control.

[0024] At the level of real-time perception of multi-source physical data, equipment operating status data is collected synchronously at millisecond intervals. Material flow status data Furthermore, a clock synchronization mechanism is employed to ensure spatiotemporal alignment of data, enabling subsequent energy efficiency characteristic calculations to be based on synchronized data reflecting the physical states of both the "energy supply side" and the "material conversion side" at the same moment. This avoids distortion of energy efficiency characteristic values ​​caused by data timing misalignment. Further, by expanding the acquisition of equipment operating parameters such as the torque of the main mill motor, the frequency of the combustion fan, the power of the cooling water pump, and the pressure of the descaling pump, as well as material flow parameters such as the instantaneous throughput obtained from roller conveyor weighing sensors and the material conveying speed obtained from speed encoders, the system can more precisely monitor energy consumption and material status at each stage of the production line, providing richer decision-making basis for subsequent refined control.

[0025] At the instantaneous energy efficiency characteristic calculation level, by using equipment operating status data... Material flow status data Input a preset energy-material coupling model and solve for the energy consumption according to... Defined instantaneous eigenvalues This transforms the energy efficiency index characterizing the energy consumption intensity per unit of effective output from a traditional ex-post statistical quantity into a real-time control variable with a definite physical unit (kJ / kg). Crucially, when the material input remains constant but the effective output decreases, this instantaneous characteristic value... The system exhibits a response characteristic that increases instantaneously in a pulse-like manner, enabling it to capture physical indicators of the equipment's operating status shifting from "load matching" to "energy efficiency mismatch" on a millisecond timescale. This allows for rapid identification and timely response to abnormal energy efficiency conditions, avoiding the technical flaw of traditional post-event statistical models where energy waste becomes a fait accompli due to response lag.

[0026] At the dual-constraint optimization decision level, by using instantaneous eigenvalues... The system compares the performance against a dynamic baseline threshold, and further verifies equipment safety indicators by invoking a process stability constraint model when energy efficiency deviations are detected. This achieves a dynamic balance between the dual objectives of "energy efficiency optimization" and "process stability assurance." Specifically, the dynamic benchmark threshold is automatically generated based on the instantaneous characteristic values ​​of qualified operating conditions within the most recent preset time window, using a sliding window statistical method and a preset multiple of the mean plus standard deviation. This allows the threshold to adaptively update with changing operating conditions, avoiding the insufficient adaptability issues caused by manually setting fixed thresholds. Simultaneously, when the equipment is detected to be in a non-steady-state condition such as switching steel grades or changing rolls, the update of the dynamic benchmark threshold is paused, preventing abnormal data under non-steady-state conditions from contaminating the benchmark threshold. The process stability constraint model prevents oxidation burn-off or incomplete heating caused by local overheating through furnace temperature uniformity variance constraints, prevents edge and center wave defects caused by rolling parameter adjustments through strip flatness constraints, and prevents winding overheating caused by continuous high-load operation of the motor through motor temperature rise rate constraints. This helps ensure that any control action does not come at the expense of product quality or equipment safety. Furthermore, by simulating and predicting the physical correction commands to be issued, and only issuing the commands when the prediction results show that all constraint indicators will not exceed the limits, a safety control mechanism of "predicting before executing" is realized.

[0027] At the closed-loop feedback control level, physical correction commands are generated based on control algorithms and sent to the process control system's actuators via industrial communication interfaces to adjust physical operating parameters, thereby improving the instantaneous characteristic values. It can automatically regress to the dynamic reference threshold range, forming a complete physical closed-loop feedback. The actuators encompass the main energy-consuming components of a hot steel rolling production line, including the gas regulating valve of the heating furnace, the frequency converter of the combustion fan, the roll gap servo mechanism of the rolling mill, and the frequency converter of the water pump in the cooling system. This enables the system to coordinate and control energy consumption throughout the entire process. The control algorithm can employ either model predictive control (MMC) or PID control. When using MMC, the optimization objective is to minimize the deviation between the instantaneous characteristic value and the dynamic reference threshold, with process stability constraints as the limiting condition. The optimal control quantity is solved in the prediction time domain, and the deviation between the measured output and the predicted output is used to correct the model state online, thereby improving the adaptability and robustness of the control.

[0028] For the heating furnace area, when the system detects that the material flow slows down due to the reduced speed of the feed rollers but the gas flow remains high, it determines that the system is in a state of high energy consumption due to dry burning. Under the premise that the surface temperature of the steel billet in the furnace has reached the upper limit of the process and the center temperature meets the process requirements, the system can automatically identify and eliminate the phenomenon of high energy consumption due to dry burning during the material waiting period by reducing the opening of the gas regulating valve in the soaking section and / or reducing the output frequency of the combustion fan frequency converter, thereby reducing the ineffective fuel consumption of the heating furnace without affecting the subsequent steel billet exit temperature and rolling quality.

[0029] In cases of material volume loss, when laser scanning data indicates that the measured volume of the intermediate billet at the roughing mill exit is less than the theoretically set threshold, suggesting irreversible volume loss due to excessive head and tail trimming or significant oxidation and burning, a physical correction command is applied to the subsequent laminar cooling process. Under the condition of meeting the coiling temperature process deviation constraints, the low-pressure cooling manifold is automatically prioritized for activation, while the inverter output frequency of the high-pressure water supply pump is reduced or shut down. This reduces the power consumption of the cooling process, physically compensating for the increased unit energy intensity caused by material loss in the preceding stages, thus achieving energy efficiency optimization across the entire process. This cross-process dynamic energy consumption compensation strategy allows this application to identify energy efficiency optimization opportunities across the entire process, rather than being limited to isolated optimization of a single process.

[0030] For the rolling zone, when abnormal high-frequency fluctuations in the main motor current of a certain stand are detected based on the main motor current data of each stand, indicating that the stand is in an inefficient and high-consumption operating zone, the automatic thickness control system dynamically distributes part of the reduction load of the stand to the adjacent high-efficiency stand while maintaining the total reduction rate. This allows the entire unit to operate in the high-efficiency range of the motor, thereby reducing the overall power consumption. At the same time, since the total reduction rate remains unchanged, the thickness specifications of the product are not affected, thus achieving a balance between energy efficiency optimization and product quality.

[0031] At the system architecture level, by deploying the feature calculation module in an edge computing unit close to the production site, data transmission latency can be reduced, meeting the real-time requirement of millisecond-level response. The hierarchical architecture design, which uses industrial communication protocols for data transmission between the multi-source sensing module and the feature calculation module, and for command issuance between the execution drive interface and the process control system via a PLC or DCS interface, allows for easy integration with existing industrial control systems, reducing implementation costs and technical risks.

[0032] In summary, this application, through the synergistic cooperation of the aforementioned technical means, has achieved a shift from "post-event statistics" to "real-time control," realizing a better dynamic balance between energy efficiency optimization and process stability, achieving cross-process energy efficiency synergistic optimization, forming a complete physical closed-loop control, and realizing adaptive regulation of production equipment.

[0033] The specification of this application contains numerous technical features distributed across various technical solutions. Listing all possible combinations of these technical features (i.e., technical solutions) would make the specification excessively lengthy. To avoid this problem, the various technical features disclosed in the above-described invention, the various technical features disclosed in the following embodiments and examples, and the various technical features disclosed in the accompanying drawings can be freely combined to form various new technical solutions (all of which are considered to have been described in this specification), unless such a combination of technical features is technically infeasible. For example, one example discloses feature A+B+C, and another example discloses feature A+B+D+E. Features C and D are equivalent technical means that serve the same function, and technically only one needs to be used; they cannot be used simultaneously. Feature E can technically be combined with feature C. Therefore, the solution A+B+C+D should not be considered as described because it is technically infeasible, while the solution A+B+C+E should be considered as described. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating an adaptive control method for steel production equipment based on instantaneous energy efficiency characteristic feedback, according to one embodiment of this application.

[0035] Figure 2 This is a schematic diagram of the system architecture of an adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback, according to another embodiment of this application. Detailed Implementation

[0036] In the following description, many technical details are presented to help the reader better understand this application. However, those skilled in the art will understand that the technical solutions claimed in this application can be implemented even without these technical details and various variations and modifications based on the following embodiments.

[0037] Explanation of some concepts: Equipment operating status data (referred to as) Energy consumption (EC) refers to the set of physical quantities characterizing the current energy consumption state of production equipment. It includes at least the instantaneous flow rate of the fuel medium and the current value of the drive motor, and can also be extended to include the torque of the rolling mill main motor, the frequency of the combustion fan, the power of the cooling water pump, and the pressure of the descaling pump, etc. These data directly reflect the energy input rate on the equipment side.

[0038] Material flow status data (denoted as) Work-in-process (WPC) refers to the set of physical quantities characterizing the current physical state of work-in-process (such as billets, intermediate billets, and strip steel). It includes at least temperature distribution and geometric dimension data obtained through infrared imaging or laser scanning, and can also be extended to include instantaneous throughput data obtained from roller conveyor weighing sensors and material transfer speed data obtained from speed encoders. These data directly reflect the effective output status on the material side.

[0039] Energy-material coupling model: This refers to a mathematical model that establishes a quantitative mapping relationship between equipment energy consumption and effective material output, with equipment operating status data as its input. Material flow status data The output is the instantaneous feature value. The model can adopt a gray box model structure based on physical mechanisms. The mechanism layer is established based on thermodynamic and rolling mechanics formulas, and the correction layer is calibrated online using historical data.

[0040] Instantaneous eigenvalue (denoted as) (): refers to a physical quantity that characterizes the energy consumption intensity per unit of effective output within the current sampling period, based on the energy consumption power of the equipment. With effective material production rate The quotient is defined in the form of kJ / kg. When the amount of material input remains constant but the effective output decreases, this instantaneous characteristic value exhibits a response characteristic of instantaneous pulse-like increase.

[0041] Dynamic baseline threshold (denoted as) ): refers to an adaptive reference value used to determine whether there is an abnormal deviation in instantaneous feature value. It is automatically generated based on the sliding window statistics of the instantaneous feature values ​​of qualified working conditions within the most recent preset time window, in the form of the mean plus a preset multiple of the standard deviation, and can be adaptively updated as the working conditions change.

[0042] Process stability constraint model: refers to the constraint index system used to verify whether the control action will destroy the stability of the production process. It includes furnace temperature uniformity variance constraint (calculated based on temperature data of multiple thermocouples in the furnace), plate flatness constraint (based on the wave index feedback of the plate shape instrument) and motor temperature rise rate constraint (calculated based on data of motor winding temperature sensor).

[0043] Equipment safety indicators (denoted as) ): refers to the state quantity that comprehensively characterizes whether the current equipment is within the allowable adjustment safety margin range, and is calculated in real time based on various constraint indicators in the process stability constraint model.

[0044] Physical correction command: refers to the control signal generated by the system based on the control algorithm, which is used to drive the actuator to adjust the physical operating parameters, including but not limited to the opening setting value of the gas regulating valve, the frequency setting value of the frequency converter, and the position setting value of the servo mechanism.

[0045] Edge computing units refer to industrial computing devices deployed close to the production site, which can complete real-time data processing and feature calculation locally, avoiding the communication delay caused by uploading massive amounts of raw data to remote servers and meeting the real-time requirement of millisecond-level response.

[0046] Process control system (PCS): refers to a computer control system used to realize the automated control of the production process. It usually includes PLC (Programmable Logic Controller) or DCS (Distributed Control System), which is responsible for receiving control commands issued by the upper system and driving the field actuators to move.

[0047] The following is a brief summary of some of the innovative aspects of this application: In summary, the technical concept of this application is to break through the inherent cognitive limitations of the existing technology that treats energy efficiency indicators only as ex-post statistics or management accounting objects, and to creatively sink the instantaneous characteristic value that represents the energy consumption intensity per unit of effective output from the management level to the control level, making it a real-time feedback state quantity with the same status as traditional process variables such as furnace temperature and rolling force. Then, within the safety margin boundary defined by the process stability constraint model, it drives physical actuators such as gas regulating valves, combustion fan frequency converters, roll gap servo mechanisms, and water pump frequency converters to perform closed-loop regulation actions.

[0048] Specifically, the technical problem addressed in this application is not simply the management issue of "how to reduce energy consumption," but rather how to achieve a dynamic balance between the seemingly mutually restrictive technical objectives of "energy efficiency optimization" and "process stability assurance" in a complex industrial process with strong coupling, multiple constraints, and time-varying disturbances, such as a hot steel rolling production line. While existing process control systems (PCS) can achieve closed-loop control of physical quantities such as temperature, pressure, and thickness, their control objectives are only directed at the physical quality indicators of the product, failing to identify suboptimal operating states such as "high quality but high energy consumption." Simultaneously, while existing energy management systems can perform energy consumption statistics and analysis, their actions lag behind the production process itself, and their output is limited to reports or alarm signals, unable to directly participate in real-time equipment control.

[0049] To address the aforementioned technical issues, this application synchronously collects equipment operating status data (including at least the instantaneous flow rate of the fuel medium and the current value of the drive motor) and material flow status data (including at least the temperature distribution, geometric dimensions, and instantaneous material throughput of the finished product) at millisecond intervals and inputs them into a pre-set energy consumption-material coupling model to calculate instantaneous characteristic values. These instantaneous characteristic values ​​are defined as the quotient of the equipment's energy consumption power and the effective material output rate, with a clear physical meaning: the energy intensity per unit of effective output, expressed in kJ / kg. Crucially, when the material input remains constant but the effective output decreases, the instantaneous characteristic value exhibits a transient pulse-like increase due to the instantaneous decrease in the denominator while the numerator remains constant or decreases with lag. This pulse-like increase is not a simple numerical fluctuation but a physical indicator of the equipment's operating status shifting from "load matching" to "energy efficiency mismatch." This application utilizes this instantaneous pulse as a signal source to trigger subsequent dual-constraint optimization decisions.

[0050] However, simply identifying energy efficiency deviations is insufficient to directly generate control commands. In the industrial setting of steel production, which has stringent requirements for temperature uniformity, plate straightness, and equipment safety, any unconstrained adjustment could lead to product quality accidents or equipment safety risks. Therefore, this application introduces a "dual constraint" decision-making mechanism: The first constraint compares instantaneous characteristic values ​​with a dynamic benchmark threshold. This dynamic benchmark threshold is not a statically fixed value, but an adaptive threshold automatically generated based on sliding window statistics of instantaneous characteristic values ​​from qualified operating conditions within the most recent preset time window, calculating the mean and standard deviation. The second constraint, after detecting energy efficiency deviations, further invokes a process stability constraint model. Based on furnace temperature uniformity variance constraints, plate straightness constraints, and motor temperature rise rate constraints, it calculates equipment safety indicators in real time. Only when the generated physical correction command, as predicted by simulation, will not cause the aforementioned constraint indicators to exceed limits, is the physical correction command issued to the actuator. This conditional control logic of "first identifying deviations, then verifying constraints, and finally implementing adjustments" enables this application to achieve a dynamic balance between the technical contradiction of "pursuing optimal energy efficiency" and "ensuring process stability." The realization of this balance depends on the synchronous calculation and collaborative judgment of instantaneous characteristic values, dynamic benchmark thresholds, and equipment safety indicators in the time dimension. The absence of any single indicator or the misalignment of the calculation sequence will lead to the failure of the decision-making mechanism.

[0051] More importantly, this application also proposes a cross-process dynamic energy consumption compensation strategy for the technical scenario of "irreversible material loss in the preceding process," which cannot be effectively addressed under the existing technical framework. When laser scanning data indicates that the measured volume of the intermediate billet at the roughing mill exit is less than the theoretically set threshold, the system identifies that the batch of material has experienced irreversible volume loss due to excessive cutting at the beginning and end or significant oxidation and burning. At this time, the instantaneous characteristic value of the unit effective output energy consumption intensity of the batch of products calculated in step 200 is... The energy consumption will inevitably increase. Traditional control methods are powerless against such irreversible losses that have already occurred. The innovation of this application lies in recognizing that although material losses in the preceding stages are irreversible, there is still room for energy efficiency optimization in the subsequent laminar flow cooling process. Under the condition of meeting the process deviation constraints of the winding temperature, the system automatically prioritizes the activation of the low-pressure cooling manifold and reduces or shuts down the inverter output frequency of the high-pressure water supply pump. By reducing the power consumption of the cooling stage, the increased unit energy intensity caused by material losses in the preceding stages is physically compensated. This cross-process energy consumption compensation mechanism allows this application to find energy efficiency optimization space throughout the entire process, rather than being limited to the isolated optimization of a single process.

[0052] In summary, the technical effects achieved in this application are not a simple summation of various technical features, but rather stem from the inherent connection and synergy between the millisecond-level impulse response characteristics of instantaneous feature values, the conditional control logic of the dual-constraint decision model, and the systematic coordination of cross-process energy consumption compensation strategies. It is precisely this multi-dimensional and multi-level coupling relationship of technical features that enables this application to achieve a technological leap from "post-event statistics" to "real-time control," from "single-process optimization" to "full-process collaboration," and from "single-objective pursuit" to "multi-objective dynamic balance."

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0054] As stated above, the inventors of this application, through long-term and in-depth research on the energy efficiency control problem of hot steel rolling production lines, have discovered that the reason why the existing technology is unable to effectively solve the above-mentioned technical problems is fundamentally due to the existence of three levels of deep-seated technical obstacles within the existing technical framework.

[0055] The inventors first conducted an in-depth analysis of the lag problem in traditional energy management models. Through statistical analysis of a large amount of production data, they discovered that the minute-level or even batch-level statistical cycles used in existing energy management systems have a significant order-of-magnitude difference from the dynamic response characteristics of the hot rolling production process. Changes in the operating conditions of a hot rolling production line often occur on a millisecond to second timescale, while by the time the energy management system detects an energy consumption anomaly, the abnormal state may have already persisted for a considerable period. The inventors realized that this mismatch in time scales is not simply a matter of technical parameters, but stems from the inherent cognitive limitation of existing technologies that position energy efficiency indicators as "management-level accounting objects." This positioning keeps energy efficiency indicators outside the real-time control loop, preventing them from participating in closed-loop regulation like traditional process variables such as furnace temperature and rolling force. Based on this understanding, the inventors creatively proposed a technical concept to "sink" energy efficiency indicators from the management level to the control level: constructing an instantaneous energy efficiency characteristic value that can be calculated on a millisecond-level cycle, making it a real-time feedback state quantity with the same status as traditional process variables.

[0056] Further research by the inventors revealed that even if energy efficiency characteristic values ​​could be calculated in real time, they could not be simply used to generate control commands, because the steel production process has extremely stringent requirements for temperature uniformity, plate flatness, and equipment safety. Through a review of historical production accident cases, the inventors discovered that in production practice, neglecting other constraints when pursuing a certain optimization goal could lead to product quality accidents or equipment failures. The inventors thus recognized a complex coupling relationship between the goals of energy efficiency optimization and process stability assurance: excessive pursuit of optimal energy efficiency could lead to problems such as excessively rapid cooling of the heating furnace causing temperature unevenness, excessive load redistribution in the rolling mill causing plate defects, or excessively rapid temperature rise due to drastic changes in motor load; while overly conservative constraints might prevent the full utilization of the energy efficiency optimization space. After repeated deliberation, the inventors proposed a "dual constraint" decision-making mechanism, which embeds a verification link of the process stability constraint model between energy efficiency deviation identification and control action execution. The control action will only be actually executed if the physical correction command to be issued will not cause the constraint indicators such as furnace temperature uniformity, plate flatness, and motor temperature rise rate to exceed the limit after simulation prediction.

[0057] More importantly, the inventors discovered a technical scenario that existing technologies are completely unable to address: the problem of increased unit energy consumption caused by irreversible material loss in upstream processes. Through long-term tracking and analysis of cutting and burning loss data in the roughing rolling process, the inventors found that material volume loss in production has a certain degree of randomness and unpredictability. Once loss occurs, it becomes a fait accompli. Traditional single-process control logic, due to its control boundary being limited to the current process, is helpless against this loss that has already occurred in the upstream process. However, after a systematic full-process energy consumption analysis, the inventors creatively realized that although material loss in upstream processes is irreversible, subsequent processes—especially laminar flow cooling processes—still have room for energy efficiency optimization. By reconstructing the cooling strategy under the condition of meeting the coiling temperature process deviation constraints, the power consumption of the cooling process can be reduced, thereby partially offsetting the increased unit energy consumption caused by upstream material loss throughout the entire process. This cross-process energy efficiency compensation mechanism breaks through the limitations of traditional single-process control and realizes synergistic energy efficiency optimization from a full-process perspective.

[0058] Based on the above in-depth research, the inventors of this application propose an adaptive control method and system for steel production equipment based on instantaneous energy efficiency characteristic feedback. The core of this technical solution lies in: constructing a real-time closed-loop feedback loop with instantaneous characteristic values ​​as the core control variables, introducing a dual decision-making mechanism based on dynamic benchmark threshold comparison and process stability constraint verification, and establishing a cross-process energy consumption dynamic compensation strategy, thereby achieving coordinated energy efficiency optimization throughout the entire process while ensuring process stability.

[0059] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the following embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Various modifications and variations can be made to this application by those skilled in the art without departing from the spirit and scope thereof.

[0060] The first embodiment of this application provides an adaptive control method for steel production equipment based on instantaneous energy efficiency characteristic feedback. Its core idea is to transform the traditional post-event energy consumption statistics model into a real-time physical closed-loop control model. Specifically, by deploying sensor groups at key nodes of the production line to collect physical data on equipment operating status and material flow status in real time, edge computing units are used to map this physical data into instantaneous characteristic values ​​representing the energy consumption intensity per unit of effective output. Under the constraints of process stability, physical correction commands are automatically generated to drive the actuators to adjust operating parameters, thereby achieving energy efficiency optimization of the production process. The following explanation uses a hot-rolled strip steel production line as an example.

[0061] System Architecture Overview Reference Figure 2 As shown in the embodiments of this application, the adaptive control system for steel production equipment includes four core functional modules: a multi-source sensing module, a feature calculation module, an optimization decision controller, and an execution drive interface. These four modules work together to form a complete physical closed-loop control architecture.

[0062] The multi-source sensing module, acting as the system's data acquisition front-end, connects to field instruments (e.g., sensor arrays) via sensor interfaces and is responsible for real-time acquisition of two types of key physical data. The first type is equipment operating status data, denoted as... The first category includes at least the instantaneous flow rate of the fuel medium and the current value of the drive motor; the second category is material flow status data, denoted as... This includes at least the temperature distribution, geometric dimensions, and instantaneous material throughput data of the product. These physical data form the basis for subsequent characteristic calculations and control decisions.

[0063] The feature calculation module is connected to the multi-source sensing module and contains an energy consumption-material coupling model. Its main function is to map the discrete physical signals acquired by the multi-source sensing module into instantaneous feature values ​​with clear physical meaning. In a preferred embodiment of this application, the feature calculation module is deployed in an edge computing unit close to the production site to reduce data transmission latency and ensure real-time control. Data transmission between the multi-source sensing module and the feature calculation module is performed using an industrial communication protocol, including but not limited to OPC-UA or Modbus protocols.

[0064] The optimization decision controller is connected to the feature calculation module, and its core function is to make optimization decisions under the constraints of the process stability constraint model. Specifically, the optimization decision controller outputs the instantaneous feature values ​​from the feature calculation module. By comparing with a dynamic benchmark threshold, when an energy efficiency deviation is detected, the process stability constraints are further verified to ensure that adjustment is permissible, and control variables are calculated based on the deviation. The process stability constraint model includes furnace temperature uniformity variance constraints, plate flatness constraints, and motor temperature rise rate constraints. These constraints help ensure that the stability of the production process is not compromised while pursuing energy efficiency optimization.

[0065] The execution drive interface connects to the optimization decision controller, responsible for converting the calculated control variables into standard PLC control commands, and sending them to the field actuators via the PLC or DCS interface. These actuators include, but are not limited to, furnace valve positioners, motor frequency converters, or hydraulic servo systems. After the actuator operates, its execution result is collected again by the sensors of the multi-source sensing module and fed back to the feature calculation module to recalculate the instantaneous feature values, thus forming a complete closed-loop control circuit.

[0066] The core advantage of the above system architecture lies in the fact that by shifting energy efficiency indicators from the management level to the control level, making them process control variables as important as temperature and pressure, real-time adaptive regulation of production equipment is achieved.

[0067] Step 100: Real-time sensing of multi-source physical data like Figure 1 As shown, step 100 is the data acquisition step of the method of this application. Its purpose is to synchronously acquire equipment operating status data at millisecond intervals using sensor groups deployed in key process sections of the steel production line (e.g., sensor groups deployed in the heating furnace, roughing mill, finishing mill, and laminar flow cooling area of ​​the steel production line). and material flow status data These two types of data respectively characterize the physical state of the energy input side and the material conversion side, providing raw data support for subsequent energy efficiency characteristic calculations.

[0068] Data on equipment operating status The data collected includes at least the instantaneous flow rate of the fuel medium and the current value of the drive motor. The instantaneous flow rate of the fuel medium is usually obtained through a vortex flow meter or mass flow meter, which directly reflects the rate of thermal energy input to the heating furnace area. The current value of the drive motor is obtained through a current transformer, which can be used to calculate the real-time power consumption of the motor, reflecting the rate of mechanical energy input to the rolling zone.

[0069] Furthermore, in order to obtain more comprehensive equipment operating status information, equipment operating status data... Further expansion includes: torque data of the rolling mill's main motor, acquired via a torque sensor, which can be used to assess the load distribution during the rolling process; frequency data of the combustion fan, read from the inverter's operating parameters, reflecting the air supply status of the combustion system; power data of the cooling water pump, obtained via a power analyzer or from the inverter, reflecting the energy consumption level of the cooling system; and pressure data of the descaling pump, acquired via a pressure sensor, reflecting the operating conditions of the descaling system. This expanded data enables the system to monitor energy consumption at each stage of the production line with greater precision.

[0070] Data on material flow status The data collected includes at least the temperature distribution and geometric dimensions of the work-in-process obtained through infrared imaging or laser scanning. Infrared imaging data, typically provided by infrared thermal imagers or infrared thermometer arrays, can acquire a two-dimensional temperature distribution field on the surface of the steel billet, which is crucial for assessing heating uniformity and predicting rolling temperature. Laser scanning data, typically provided by laser diameter gauges or laser profilometers, can acquire the geometric dimensions of the work-in-process, such as thickness and width, in real time. These data directly reflect the deformation state of the material.

[0071] Furthermore, material flow status data It can also be expanded to include: instantaneous throughput based on roller conveyor weighing sensors, which reflects the mass of material passing through a specific station per unit time and is a key parameter for calculating the effective output rate; and material conveying speed based on speed encoders, which, combined with geometric data, can be used to calculate the volumetric flow rate of the material.

[0072] It is particularly important to note that, since the sampling periods of the various sensors mentioned above may differ, this application employs a clock synchronization mechanism to ensure spatiotemporal alignment of the data. In a preferred embodiment, the data from each sensor can be clock-synchronized using a precision time protocol commonly used in the industrial field, and the sampling period can be configured within the range of tens to hundreds of milliseconds depending on the control accuracy requirements. For signals with different sampling rates, the system uses a sliding window interpolation algorithm or a timestamp alignment mechanism for time alignment, which helps ensure that the data input to the feature calculation module remains consistent in the time dimension.

[0073] Step 200: Instantaneous Energy Efficiency Characteristics Calculation Step 200 is the feature calculation step of the method of this application. Its purpose is to utilize the edge computing unit to process the equipment operating status data collected in step 100. Material flow status data Input a preset energy consumption-material coupling model and calculate the instantaneous characteristic value representing the energy consumption intensity per unit of effective output within the current sampling period. .

[0074] The instantaneous feature value The physical meaning of this characteristic value is: the energy consumed in transforming a unit mass of material from its current state to the next process state within the current sampling period. This characteristic value is an engineering parameter with a definite physical unit, which is kJ / kg. Essentially, it characterizes the degree to which the current operating point of the equipment deviates from the optimal energy efficiency curve. When the equipment is in a high-efficiency operating state, this characteristic value approaches the theoretical minimum; when the equipment is in an inefficient and high-consumption state, this characteristic value will increase significantly.

[0075] Specifically, the instantaneous feature value Solve using the following formula: (Formula 1) In Formula 1, These are instantaneous characteristic values, representing the energy intensity per unit of effective output; This represents the power consumption of the device during the current sampling period. This represents the effective material output rate within the current sampling period. The calculation methods for these two parameters are explained below.

[0076] Equipment energy consumption power The total energy consumption power is obtained by summing the thermal power on the fuel side and the electrical power on the power side. More specifically, the thermal power on the fuel side equals the instantaneous flow rate of the fuel medium multiplied by the fuel calorific value, and then multiplied by the combustion efficiency coefficient; the electrical power on the power side is obtained by multiplying the current and voltage of the drive motor, or by reading the power signal directly output by the frequency converter. The sum of these two values ​​yields the total energy consumption power, expressed in kW. It should be noted that in practical applications, the fuel calorific value can be measured in real time using an online calorimeter, or a standard value can be obtained by looking up a table based on the fuel type; the combustion efficiency coefficient can be calibrated based on the design parameters of the heating furnace and historical operating data.

[0077] Effective material production rate The calculation relies on material flow status data. Specifically, the mass flow rate is obtained by multiplying the roller conveyor weighing signal and the speed measurement signal, or by multiplying the real-time cross-sectional area of ​​the work-in-process obtained by laser diameter measurement with the material transport speed and material density. This mass flow rate is the effective material output rate, and its unit is kg / s.

[0078] A key technical feature of this application is: when the material input remains constant but the effective output decreases, the instantaneous characteristic value... It exhibits a transient, pulse-like increase. This pulse-like response characteristic allows the system to detect abnormal changes in equipment operating status on a millisecond timescale. For example, when the material's time in the furnace is prolonged due to a slowdown in the preceding process, while the gas supply remains constant, the effective material output rate... Decrease, according to Formula 1, instantaneous eigenvalue When the value increases, the system can identify the "high consumption due to idling" state. Similarly, when excessive cutting at the beginning and end of the rolling process leads to increased material volume loss, the effective output decreases, and the instantaneous characteristic value will also show a pulse-like increase, which the system can use to trigger subsequent compensation control.

[0079] Regarding the structure of the energy consumption-material coupling model, this application adopts a gray box model based on physical mechanisms. The input layer of this model receives the real-time signal vector from step 100. The mechanism layer establishes the furnace heat balance equation based on the first law of thermodynamics and the deformation work equation based on rolling mechanics formulas. The correction layer uses historical production data to perform online corrections on the thermal efficiency coefficient and friction coefficient in the model. The output layer outputs instantaneous characteristic values. This gray box model structure balances the interpretability of physical mechanisms with the adaptability of data-driven approaches. In practical applications, the specific parameters of the model can be calibrated and optimized based on production data for different steel grades and specifications.

[0080] Step 300: Dual-Constraint Optimization Decision Step 300 is the optimization decision-making step of the method in this application. Its core function is to make energy efficiency optimization decisions while ensuring process stability. This step introduces a "dual constraint" mechanism: the first constraint is the comparison between instantaneous feature values ​​and dynamic benchmark thresholds, used to identify energy efficiency anomalies; the second constraint is the verification of the process stability constraint model, used to ensure that adjustment actions will not disrupt the stability of the production process. This dual constraint mechanism is one of the core innovations of this application, effectively resolving the technical contradiction that simply pursuing energy efficiency optimization may lead to a decrease in process stability.

[0081] Step 310: Dynamic benchmark threshold comparison First, the instantaneous feature values ​​obtained in step 200 are processed. The energy efficiency deviation is determined by comparing the data with a dynamic benchmark threshold. This dynamic benchmark threshold is not a static value that is fixed manually, but rather an adaptive threshold that is automatically generated and dynamically updated based on historical stable operating data.

[0082] Specifically, the dynamic benchmark threshold is generated as follows: the system generates the instantaneous characteristic values ​​of qualified operating conditions within the most recent preset time window. Perform sliding window statistics and calculate the mean within the window. with standard deviation Then add the mean to the standard deviation by a preset factor. This serves as the upper limit of the threshold. The calculation process can be expressed by the following formula: (Formula 2) (Formula 3) (Formula 4) In the above formula, The mean of the instantaneous feature values ​​within the sliding window; The standard deviation of the instantaneous eigenvalues ​​within the sliding window; This represents the number of sampling points within the sliding window; This is a configurable sensitivity coefficient; This represents the upper limit of the dynamic baseline threshold. Sensitivity coefficient. The value of determines the system's sensitivity to energy efficiency deviations: The smaller the value, the more sensitive the system is, but it may generate more false triggers; A higher value results in a more robust system, but may miss some abnormal operating conditions. In practical applications, the value can be adjusted according to the specific characteristics of the production line and the control precision requirements. The value of k depends on the production line's tolerance for energy efficiency deviations, and typically ranges from 2 to 4.

[0083] It is important to note that the generation of dynamic benchmark thresholds also includes logic for identifying and eliminating unsteady operating conditions. When the system detects that the equipment is in an unsteady operating condition, such as switching steel grades or changing rolls, it pauses the updating of dynamic benchmark thresholds to prevent abnormal data under unsteady operating conditions from contaminating the benchmark thresholds. The identification of unsteady operating conditions can be based on operating condition switching signals issued by the production scheduling system, or it can be automatically identified based on the statistical characteristics of physical data (such as variance mutations, mean shifts, etc.).

[0084] Step 320: Process stability constraint verification When step 310 detects the instantaneous feature value When the threshold exceeds the upper limit of the dynamic baseline threshold (i.e., higher than or higher than the upper limit of the dynamic baseline threshold), the system does not immediately generate an adjustment command. Instead, it further calls the process stability constraint model to calculate the equipment safety index in real time based on the data collected in step 100. Verify whether the current device is within the allowable adjustment margin (i.e., all constraint indicators are within limits / meet preset limits).

[0085] The process stability constraint model includes the following three types of constraint indices: The first type is the furnace temperature uniformity variance constraint. This constraint calculates the temperature variance based on temperature data from multiple thermocouples within the furnace. And determine whether the temperature variance is less than the upper limit of variance set in the process specification. The physical significance of furnace temperature uniformity constraints lies in preventing excessively high or low local temperatures from accelerating oxidation and burning of the steel billet or causing incomplete heating, thereby affecting the quality of subsequent rolling. Only when the temperature distribution within the furnace is sufficiently uniform can the system allow for a reduction in the gas supply to the heating furnace.

[0086] The second type is the plate shape straightness constraint. This constraint is based on the wave index obtained from the real-time feedback signal of the plate shape meter. And determine whether the wave index is within the preset controlled range. The physical significance of strip flatness constraints lies in preventing strip shape defects such as edge waviness and center waviness caused by adjustments to rolling parameters, thereby affecting finished product quality and coiling stability. The system only allows adjustments to the mill roll gap or load distribution when the strip shape index is within the preset controlled range.

[0087] The third type is the motor temperature rise rate constraint. This constraint calculates the temperature rise rate based on data from motor winding temperature sensors. The system then determines whether the temperature rise rate is below a safe threshold. The physical significance of the motor temperature rise rate constraint is to prevent the windings from overheating and burning out due to continuous high-load operation of the motor. Only when the motor temperature rise rate is within a safe range is the system allowed to redistribute the motor load.

[0088] The above three types of constraint indicators together constitute the equipment safety indicators. The system employs a predictive verification mechanism when making adjustment decisions: firstly, it simulates and predicts the physical correction command to be issued, evaluating the changing trends of various constraint indicators after the command is executed; only when the prediction results indicate that all constraint indicators will not exceed limits is the physical correction command actually issued to the actuator. This predictive verification mechanism effectively resolves the technical contradiction that "pursuing energy efficiency optimization alone may damage process stability." The simulation prediction can be implemented using a simplified mechanistic model or an empirical model trained based on historical data, with the core purpose of predicting the consequences of control actions before execution.

[0089] Furthermore, when the instantaneous eigenvalue When optimization is needed but the process stability constraint margin of the current process is insufficient, the system's decision logic does not simply abandon optimization, but instead turns to searching for other adjustable processes or parameters. This "cross-process optimization" decision strategy is one of the important innovations of this application, and its specific implementation will be described in detail in step 400.

[0090] Step 400: Closed-loop feedback control Step 400 is the closed-loop control step of the method in this application. When step 300 determines the instantaneous characteristic value... Abnormal deviation and equipment safety indicators When safety margin requirements are met, the system generates physical correction commands for the actuators of the production equipment based on the control algorithm, and sends them to the process control system through the industrial communication interface. The actuators then adjust their physical operating parameters to ensure that the instantaneous characteristic values ​​are met. The dynamic baseline threshold range is regressed to form a closed-loop feedback.

[0091] Step 410: Actuator and Control Algorithm Regarding the actuators, this application involves at least one of the following: a gas regulating valve for a heating furnace, used to adjust the valve opening to control the fuel supply; a frequency converter for a combustion fan, used to adjust the fan frequency to control the combustion air volume; a roll gap servo mechanism for a rolling mill, used to adjust the roll gap setpoint to control the reduction; and a water pump frequency converter for a cooling system, used to adjust the pump's operating frequency to control the cooling water flow. These actuators cover the main energy-consuming stages of a hot-rolled steel production line. Through coordinated control of these actuators, energy efficiency optimization throughout the entire process can be achieved.

[0092] Regarding the control algorithm, this application may employ at least one of model predictive control (MMC) or PID control. In a preferred embodiment, when using MMC, the system uses instantaneous characteristic values... The optimization objective is to minimize the deviation from the dynamic reference threshold. The optimal control quantity is solved in the prediction time domain using constraints from the process stability constraint model, including furnace temperature uniformity variance, plate flatness, and motor temperature rise rate. Furthermore, the system uses the deviation between the measured and predicted outputs to perform online corrections to the model state, thereby improving the model's prediction accuracy and control robustness. Specific parameters in the prediction and control time domains can be configured according to the dynamic characteristics of the controlled object and the control accuracy requirements.

[0093] The generation and issuance of physical correction instructions follow this process: First, the control algorithm calculates the increment of the control variable based on the deviation of the instantaneous characteristic value and the constraint conditions; then, the increment of the control variable is added to the current setpoint to obtain a new setpoint; finally, the new setpoint is converted into a standard PLC control instruction format and issued to the corresponding actuator through the industrial communication interface. After the actuator adjusts the physical operating parameters according to the received instructions, the sensor group in step 100 will re-collect equipment operating status data and material flow status data, and the feature calculation module in step 200 will recalculate the instantaneous characteristic value, thus forming a complete closed-loop feedback loop.

[0094] The control strategy implementation for step 400 will be explained in detail below using three specific application scenarios.

[0095] Step 420: Control strategy for the heating furnace area This embodiment describes an adaptive control strategy for the "high consumption due to dry burning" state in the heating furnace area, which corresponds to solving the problem of ineffective combustion caused by the slowdown of the previous process.

[0096] When step 100 detects a decrease in the speed of the furnace feed rollers, resulting in a slower material flow, but the gas flow meter reading remains high, step 200 calculates the instantaneous characteristic value. The value will increase. This is because the denominator (effective material output rate) decreases while the numerator (energy consumption power) remains unchanged. According to Formula 1, the quotient must increase. The system classifies this state as "high-consumption, no-load combustion," meaning that the furnace maintains high-level combustion even with a reduced material flow rate, resulting in a large amount of heat energy being lost as waste heat from the flue gas.

[0097] After proceeding to step 300, the system invokes the process stability constraint model for verification. Specifically, the system checks whether the surface temperature of the billet in the furnace has reached the upper limit set by the process specification, and simultaneously checks whether the center temperature of the billet has met the minimum requirement of the process. If the surface temperature has reached the upper limit and the center temperature meets the requirements, it is determined that there is a margin for cooling under the current operating conditions, the stability constraint verification passes, and the fuel supply to the heating furnace can be reduced. The determination of the billet center temperature can be achieved based on online calculation of the furnace thermal model or estimation using empirical formulas.

[0098] The physical correction command generated in step 400 includes a combination of one or more of the following actions: reducing the opening of the gas regulating valve in the soaking section, and / or reducing the output frequency of the combustion blower inverter, thereby reducing heat loss carried away by excess air. This combination of regulating actions can effectively reduce the supply of ineffective calorific value while maintaining the furnace material temperature at the lower limit of the process requirements.

[0099] The technical effect of this control strategy is that during the material waiting period (such as the waiting state caused by the slowdown of the previous process), the system can automatically identify and eliminate the phenomenon of "empty burning and high consumption", reduce the ineffective fuel consumption of the heating furnace, and at the same time, it does not affect the subsequent billet exit temperature and rolling quality.

[0100] Step 430: Dynamic Energy Consumption Compensation Strategy Across Processes This embodiment describes one of the core innovations of this application: a dynamic energy consumption compensation strategy across processes to address material volume loss. This strategy addresses the problem of how to physically compensate for irreversible material loss in subsequent processes through energy-saving operations, thereby maintaining the overall energy efficiency optimization target.

[0101] The specific scenario is as follows: When the laser scanning data in step 100 indicates that the measured volume of the intermediate billet at the roughing mill exit is less than the theoretically set threshold, the system identifies that irreversible volume loss has occurred in this batch of material. The causes of volume loss may include: the amount of head and tail trimming exceeding expectations, the degree of oxidation and burn-off exceeding the design value, etc. Regardless of the specific cause, the physical consequences of volume loss are clear: since the preceding stages have already consumed corresponding energy for heating and roughing, the effective output material quality is reduced. Therefore, the instantaneous characteristic value of the unit effective output energy consumption intensity of this batch of product calculated in step 200 will decrease. It will inevitably rise.

[0102] Traditional control methods are powerless against such irreversible losses that have already occurred, and can only perform statistical analysis afterward. However, the innovation of this application lies in recognizing that although material losses in upstream processes are irreversible, there is still room for energy efficiency optimization in downstream processes. By implementing energy-saving operations in downstream processes, the increase in unit energy consumption caused by upstream losses can be partially offset.

[0103] Specifically, the physical correction command applies to the subsequent laminar flow cooling process. Located after finishing rolling and before coiling, the laminar flow cooling process primarily cools the finished strip from the finishing rolling end temperature to the coiling temperature. A key constraint of the cooling process is that the coiling temperature must be controlled near the target value set in the process specifications to ensure that the metallographic structure and mechanical properties of the strip meet product standards.

[0104] Under the condition of meeting the winding temperature process deviation constraints, the system automatically reconfigures the cooling manifold activation strategy: prioritizing the activation of the low-pressure section cooling manifold and reducing or shutting down the inverter output frequency of the high-pressure water supply pump. The physical principle behind this strategy is that the water pressure in the low-pressure section cooling manifold is lower, resulting in lower pump power consumption per unit time; while the high-pressure water supply pump is the main power-consuming equipment in the cooling system, reducing its operating frequency can significantly reduce energy consumption. Through the reconfiguration of the cooling strategy, the energy consumption of the cooling process is reduced while achieving the same cooling effect.

[0105] From the perspective of overall energy efficiency optimization, the technical effect of this cross-process compensation strategy is that it physically compensates for the increased unit energy consumption caused by material loss in the preceding processes by reducing the electrical energy consumption in the cooling process, thus achieving overall energy efficiency optimization. It should be noted that this compensation is carried out under the premise of meeting product quality constraints (winding temperature deviation) and will not adversely affect the quality of the final product.

[0106] Step 440: Dynamic Load Allocation Strategy for Rolling Mill This embodiment describes an energy efficiency optimization control strategy for the rolling zone. Its core idea is to dynamically distribute the load between the stands so that the entire unit operates in the high-efficiency range of the motor.

[0107] In a continuous rolling mill, the main motors of each stand undertake different proportions of the rolling task. Due to various disturbances during the rolling process, the main motors of some stands may occasionally deviate from the high-efficiency operating range, exhibiting an inefficient and high-consumption operating state. This state is physically characterized by: abnormally high-frequency fluctuations in the main motor current and an extremely low energy efficiency ratio (the ratio of output rolling work to input electrical work).

[0108] Step 100 monitors the current waveform based on the current data of the main motors in each rack. When abnormal high-frequency fluctuations are detected in the main motor current of a rack, the energy efficiency ratio calculation result in step 200 will show that the rack is in an inefficient and high-consumption operating zone. The identification of abnormal high-frequency fluctuations can be achieved based on the statistical characteristics of the current signal (such as variance and peak value) or spectrum analysis.

[0109] After constraint verification in step 300, step 400 generates a load redistribution command using the automatic thickness control system. The constraint of this command is to maintain a constant total reduction rate, meaning the sum of reductions across all racks in the unit remains constant, to ensure the final product thickness specifications are not affected. Under this constraint, the system dynamically distributes a portion of the reduction load from racks operating in inefficient zones to adjacent high-efficiency racks.

[0110] The technical benefits of this strategy are: by dynamically balancing the load between racks, the main motors of each rack operate near their high-efficiency range, reducing the overall power consumption of the unit. Simultaneously, since the total reduction ratio remains constant, the product thickness specifications are unaffected, achieving a balance between energy efficiency optimization and product quality.

[0111] In a second embodiment of this application, an adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback is provided to implement the above-mentioned method. For example... Figure 2 As shown, the system comprises four core components: a multi-source perception module, a feature calculation module, an optimization decision controller, and an execution drive interface. The connections and coordination between these modules are as follows.

[0112] The multi-source sensing module is used to connect to field instruments (e.g., sensor arrays) via sensor interfaces to collect equipment operating status data in real time. and material flow status data This module is equivalent to the system's "perception layer," serving as the data source for the entire closed-loop control system. This includes at least the instantaneous flow rate of the fuel medium and the current value of the drive motor. This includes at least the temperature distribution, geometric dimensions, and instantaneous material throughput data of the product.

[0113] The feature calculation module is connected to the multi-source sensing module and configured to run an energy consumption-material coupling model, mapping the data collected by the multi-source sensing module into instantaneous feature values. This module is equivalent to the system's "computation layer," responsible for transforming raw physical data into feature quantities with controllable significance. In a preferred embodiment of this application, the feature calculation module is deployed in an edge computing unit close to the production site to meet the real-time requirement of millisecond-level response.

[0114] The optimization decision controller is connected to the feature calculation module and is configured to, under the constraints of the process stability constraint model, calculate the instantaneous feature values. The control variables are calculated based on the deviation from the instantaneous characteristic values ​​by comparing with a dynamic benchmark threshold. This module is equivalent to the system's "decision layer" and is the core of the "dual constraint" decision-making mechanism. Among them, the process stability constraint model includes furnace temperature uniformity variance constraints, plate flatness constraints, and motor temperature rise rate constraints.

[0115] The execution drive interface connects to the optimization decision controller and is used to convert control variables into standard PLC control instructions, which are then output to the furnace valve positioner, motor frequency converter, or hydraulic servo system to adjust the physical operating parameters of the equipment. This module is equivalent to the system's "execution layer," responsible for transforming abstract control variables into concrete physical actions.

[0116] The above four modules form a complete closed loop: the execution result is fed back to the feature calculation module through the multi-source sensing module, enabling the system to continuously monitor the control effect and make dynamic corrections.

[0117] In terms of communication architecture, the multi-source sensing module and the feature calculation module can use industrial communication protocols for data transmission, while the execution drive interface and the process control system can exchange commands via a PLC or DCS interface. This layered architecture design allows this application to be easily integrated with existing industrial control systems, reducing implementation costs and technical risks.

[0118] The first embodiment is a method embodiment corresponding to this embodiment. The technical details in the first embodiment can be applied to this embodiment, and the technical details in this embodiment can also be applied to the first embodiment.

[0119] Summary of technical effects The above-described technical solution achieves the following technical effects: First, it achieves a shift from "post-event statistics" to "real-time control." Traditional energy management methods rely on data collection and statistical analysis after production ends, which has a significant lag. This application, through millisecond-level real-time data acquisition and edge computing, can respond instantly when energy efficiency deviations occur, preventing energy waste from becoming a fait accompli.

[0120] Secondly, a better dynamic balance is achieved between energy efficiency optimization and process stability. By introducing a "dual constraint" decision-making mechanism, this application, while pursuing energy efficiency optimization, always places process stability in the position of a constraint condition, ensuring that any adjustment action will not lead to a decline in product quality or equipment safety risks.

[0121] Third, it achieves cross-process energy efficiency synergy optimization. Through the "cross-process energy consumption compensation strategy", this application can find the energy efficiency optimization space throughout the entire process. Even if irreversible material loss occurs in a certain process, it can be compensated by energy-saving measures in subsequent processes, thereby achieving the overall energy efficiency of the system.

[0122] Fourth, a complete physical closed-loop control is formed. The output of this application is not a report or alarm, but a control command that directly drives physical actuators such as valves, frequency converters, and servo mechanisms, so that instantaneous characteristic values ​​can automatically return to the reference range, realizing adaptive control of production equipment.

[0123] Furthermore, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the various steps of the above-described method. The computer-readable storage medium may be flash memory, a hard disk, an optical disk, or other non-volatile storage media.

[0124] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

[0125] Real-time verification of carbon inventory data based on instantaneous energy efficiency characteristics feedback The technical solution of this application can also be applied to the scenario of real-time carbon accounting inventory data verification in the steel production process. Since there is a definite physicochemical relationship between carbon emissions and energy consumption in the steel production process, the instantaneous energy efficiency characteristic value constructed in this application can serve as the basis for real-time physical verification of carbon accounting inventory data.

[0126] Specifically, carbon emissions in steel production mainly originate from two physical processes: direct emissions from fuel combustion and indirect emissions from purchased electricity. For direct emissions, carbon in the fuel reacts chemically with oxygen during combustion to produce carbon dioxide, and there is a deterministic physical relationship between the amount of carbon emitted and the amount of fuel consumed, based on stoichiometry. For indirect emissions, there is also a quantifiable correspondence between electricity consumption and the average carbon emission factor of the power grid.

[0127] Based on the above physicochemical relationships, this application can obtain the instantaneous eigenvalues ​​calculated in step S2. Further converted into instantaneous carbon emission intensity The specific conversion method is as follows: in, The instantaneous flow rate of the fuel medium collected in step S1; The carbon emission factor of this fuel is determined by the carbon content and combustion efficiency of the fuel, and its unit is 1 / 2 ppm. or ; The measured power of the drive motor collected in step S1; The carbon emission factor of the power grid, its unit is ; and This is a conversion factor for units of measurement.

[0128] The above instantaneous carbon emission intensity The calculation relies entirely on the physical sensor data collected in real time by the multi-source sensing module in step S1, thus providing a technical basis for real-time verification of carbon accounting inventory data.

[0129] In carbon inventory data verification applications, companies typically rely on fixed emission factors and total emission data within a statistical period when reporting carbon emissions. However, due to dynamic changes in production conditions, fluctuations in equipment operating efficiency, and differences in material properties, static accounting methods based on fixed parameters may deviate from actual emissions. This application, through millisecond-level physical data acquisition and instantaneous feature calculation, can provide dynamic carbon emission data based on actual operating conditions, thereby enabling real-time physical verification of the data in the static accounting inventory.

[0130] The specific verification mechanism is as follows: the system will calculate the instantaneous carbon emission intensity from the edge computing unit. Integrating over a preset statistical period yields the cumulative carbon emissions based on measured physical data for that period. Simultaneously, the declared carbon emissions for that period, calculated based on standard emission factors and statistical data, are extracted from the carbon accounting inventory. By comparison and deviation rate The system can identify whether there are significant deviations in the carbon accounting inventory data.

[0131] When deviation rate When the preset threshold is exceeded, the system automatically generates a verification anomaly flag and traces the source of the deviation based on the detailed physical data collected in step S1. Possible sources of deviation include: the difference between the actual calorific value and the nominal calorific value of the fuel (which can be identified by comparing the measured data of the online calorific value meter with the nominal value), the difference between the actual operating efficiency and the design efficiency of the motor (which can be identified by comparing the torque sensor and power data), and the difference between the actual amount of material processed and the declared amount (which can be identified by comparing the data of the roller conveyor weighing sensor with the production records), etc.

[0132] The technical feature of this verification mechanism is that it is entirely based on the real-time data collected by the physical sensor group deployed in step S1 of this application, and is solved by the edge computing unit in step S2, utilizing the instantaneous energy efficiency characteristic value. The same data acquisition channels and computing architecture are used. Therefore, the verification results of the carbon accounting inventory data directly reflect the actual physical operating status of the production equipment, rather than being purely numerical calculations based on manually set accounting rules.

[0133] Furthermore, when the system identifies deviations in the carbon accounting inventory data through the aforementioned verification mechanism, it can feed back the verification results to the optimization decision controller in step S4. Under the premise of satisfying process stability constraints, the optimization decision controller can generate physical correction commands. By adjusting actuator parameters such as the opening degree of the gas regulating valve and the frequency of the frequency converter, the actual carbon emission intensity can be made closer to the target value, achieving closed-loop control of carbon emissions. This carbon emission control mechanism based on real-time physical measurement shares the same technical architecture and control logic as the energy efficiency closed-loop control in the main technical solution of this application.

[0134] Further explanation of key technical features To facilitate those skilled in the art in implementing the technical solutions of this application, further explanations and exemplary implementations of several key technical features involved in the foregoing embodiments are provided below. It should be noted that the specific formulas and parameters given below are merely illustrative examples, and those skilled in the art can make adaptive adjustments based on the specific circumstances of the actual production line without departing from the scope of protection of this application.

[0135] 1. Regarding instantaneous eigenvalues Specific calculations As mentioned earlier, instantaneous eigenvalues According to the formula Solution. To enable those skilled in the art to implement this method, exemplary calculation methods for the numerator and denominator are given below.

[0136] Regarding equipment energy consumption power The specific calculation can be performed using the following formula: in, The instantaneous flow rate of the fuel medium can be obtained in real time through a vortex flow meter or a mass flow meter; The lower heating value of the fuel can be measured in real time using an online calorific value meter or obtained by looking up a standard value in a table according to the fuel type. The actual power of the drive motor can be calculated by multiplying the current and voltage or read directly from the power signal output by the frequency converter. The effective utilization coefficient of thermal power is used to calculate the proportion of fuel combustion heat actually used for heating materials. Its value ranges from 0.5 to 0.9, and the specific value can be calibrated based on the design parameters and historical operating data of the heating furnace. This is the effective utilization factor of electrical power, which is usually taken as 1.0.

[0137] Regarding the effective material production rate The specific calculation can be performed using the following formula: in, The density constant of the material being processed is typically taken as 7850 kg / m³ for steel. 3 ; Real-time cross-sectional area of ​​work-in-process obtained by laser diameter measuring instrument or laser profilometer; This refers to the real-time material transfer speed obtained from the speed encoder. All of the above parameters can be obtained from the material flow status data collected in step S1. It is obtained or calculated from [the data].

[0138] It should be noted that the terms "instantaneous characteristic value" and "energy consumption characteristic value per ton" in this application have the same meaning, both representing the energy intensity per unit of effective output, with the unit being kJ / kg. The two expressions can be used interchangeably in this application.

[0139] 2. Specific implementation of the energy consumption-material coupling model As mentioned above, this application uses a gray box model based on physical mechanisms as the energy consumption-material coupling model. To enable those skilled in the art to implement this model, exemplary forms of the heat balance equation and deformation work equation are given below.

[0140] The furnace heat balance equation can be established in the following form based on the first law of thermodynamics: in, The input thermal power corresponds to the heat released during fuel combustion; The specific heat capacity of the steel billet can be obtained from a table based on the steel grade or calculated using an empirical formula. The mass flow rate of the steel billet; and These are the billet's exit temperature and entry temperature, respectively, which can be obtained using an infrared thermometer or thermocouple. The heat loss due to furnace heat dissipation and flue gas carryover can be estimated using a furnace thermal model or calibrated based on historical data. This equation reflects the energy conservation relationship in the heating furnace region and is the physical basis for calculating the energy efficiency characteristic value of the heating process.

[0141] The equation for the work done in rolling deformation can be established in the following form based on the formulas of rolling mechanics: in, This refers to the rolling power. The width of the strip steel plate; The contact arc length between the roll and the strip can be calculated based on the roll radius and the reduction amount. The average rolling pressure can be calculated using a rolling force model or measured by the rolling mill's force measuring device. For rolling speed; This represents mechanical efficiency, used to calculate the proportion of the motor's input power actually used for rolling deformation. This equation reflects the energy conversion relationship in the rolling process and is the physical basis for calculating the energy efficiency characteristic value of the rolling process.

[0142] In practical applications, the energy-material coupling model calculates theoretical energy consumption by collecting real-time data and substituting it into the aforementioned equations. This data is then combined with measured energy consumption for online correction, resulting in highly accurate instantaneous energy efficiency characteristic values. Specifically, the model's correction layer uses historical production data to calibrate model parameters such as the thermal efficiency coefficient and friction coefficient online, enabling the model output to adapt to different steel grades and specifications under various production conditions. Furthermore, the online calibration of the correction layer can be achieved using recursive least squares or Kalman filtering algorithms. Taking the online correction of the thermal efficiency coefficient as an example, the system periodically compares the theoretically calculated energy consumption value with the measured energy consumption value, updating the estimated value based on the deviation between the two, thus gradually approximating the actual operating conditions. This online correction mechanism allows the energy-material coupling model to adapt to the effects of slowly changing factors such as equipment aging and ambient temperature variations.

[0143] 3. Regarding equipment safety indicators Calculation As mentioned earlier, the process stability constraint model includes three types of constraint indicators: furnace temperature uniformity variance constraint, plate flatness constraint, and motor temperature rise rate constraint. This is to enable those skilled in the art to implement the equipment safety indicators. The calculation is illustrated below with an exemplary implementation.

[0144] First, calculate the safety margin of each constraint index relative to its limit. For the furnace temperature uniformity variance constraint, the safety margin is... Defined as: in, This is the real-time temperature variance calculated based on temperature data from multiple thermocouples inside the furnace. The upper limit of temperature variance set for the process specification.

[0145] For plate straightness constraints, safety margin Defined as: in, The wave index is obtained based on the real-time feedback signal from the plate shape analyzer. This represents the controlled range of the wave index.

[0146] For motor temperature rise rate constraints, safety margin Defined as: in, The real-time temperature rise rate is calculated based on data from the motor winding temperature sensor. This is the safe threshold for the rate of temperature rise.

[0147] Based on this, equipment safety indicators It can be defined as the minimum value of each of the above margin measures: when When all constraints are within the safety margin, energy efficiency adjustments are permitted; when When this occurs, it indicates that at least one constraint indicator has exceeded or is close to its limit, and adjustments that could further worsen that constraint indicator are not permitted. Through the above calculation method, the "safety margin range" mentioned in step S3 has an implementable decision path.

[0148] 4. Generation and prediction verification of physical correction commands To enable those skilled in the art to specifically implement the physical correction instruction generation and prediction verification described in step S4, an exemplary implementation method is given below.

[0149] Regarding the generation of physics correction commands, the control algorithm is based on instantaneous feature values. With dynamic benchmark threshold The deviation between them is used to calculate the increment of the control variable. Specifically, when determining... When the threshold exceeds the upper limit of the dynamic baseline threshold, the control algorithm adjusts the control algorithm based on the deviation. Given process stability constraints, solve for the optimal control variable increment that makes the deviation approach zero. The increment of the control variable This could be the opening increment of a gas regulating valve, the frequency setting increment of a frequency converter, or the position setting increment of a servo mechanism. The control variable increment... The new setting value is obtained by superimposing it on the current setting value; then the new setting value is converted into the standard PLC control instruction format to form the physical correction instruction.

[0150] Regarding the prediction and verification mechanism, before the physical correction command is actually issued, the system verifies the candidate commands. Perform simulation prediction. Specifically, apply candidate instructions to a simplified mechanism model or an empirical model trained on historical data corresponding to the controlled object, predict the changing trends of various constraint indices in the prediction time domain after the instruction is executed, and obtain the predicted values. , , Calculate the forecast margin based on the predicted values. , , Thus, the predicted safety indicators are obtained. .

[0151] Only when In other words, only when the prediction results show that all constraint indicators will not exceed the limits within the prediction time domain will the candidate instructions be converted into standard PLC control instructions and sent to the process control system via the industrial communication interface to drive the actuator. If the prediction verification fails (i.e., If the current settings remain unchanged, the optimization decision controller will regenerate candidate instructions that meet the constraints. Through the above "predict first, then execute" verification process, it can be ensured that energy efficiency adjustment actions are always controlled within the safety boundaries defined by the process stability constraint model, effectively resolving the technical contradiction that simply pursuing energy efficiency optimization may damage process stability.

[0152] 5. Supplementary explanations regarding certain details in the foregoing embodiments To further clarify some of the process logic and judgment criteria involved in the foregoing embodiments, and to facilitate implementation by those skilled in the art, the following exemplary supplementary explanation is provided.

[0153] about Figure 1 The decision branch in step 300, such as Figure 1 As shown, the judgment step 300 has two branch exits. When the judgment result is "yes" (i.e., the instantaneous characteristic value exceeds the dynamic reference threshold, and the process stability constraint allows adjustment), the process proceeds to step 400 to execute closed-loop feedback control. When the judgment result is "no" (i.e., the instantaneous characteristic value does not exceed the dynamic reference threshold, or although it exceeds it, the process stability constraint does not allow adjustment), the system does not generate a physical correction command, keeps the current equipment operating parameters unchanged, and returns to step 100 to continue data acquisition and monitoring for the next sampling cycle.

[0154] Regarding the "theoretical threshold" mentioned in step 430, its specific determination method is as follows: Before the billet enters the furnace, based on the original dimensions of the billet (length × width × thickness) and the target intermediate billet specifications, and based on the principle of constant volume during rolling deformation, the theoretical volume Vtheory of the intermediate billet exiting the roughing mill is calculated. Considering the reasonable volume loss rate ηloss caused by head and tail cutting and oxidation loss under normal production conditions (its typical range is 2% to 5%, and the specific value can be determined according to the steel grade and historical statistical data), the theoretical threshold is set as follows: When laser scanning data indicates the measured volume of the intermediate billet... At that time, the system determined that the batch of materials had experienced irreversible volume loss beyond the normal range.

[0155] Regarding the “inefficient and high-consumption operating zone” mentioned in step 440, the quantitative determination method is as follows: Calculate the instantaneous energy efficiency ratio ηstand(t) of the main motor of each frame within the current sampling window (defined as the ratio of the rolling deformation work output by the frame to the input electrical power). When the instantaneous energy efficiency ratio ηstand(t) of a frame is lower than the average energy efficiency ratio of the frame under stable operating conditions minus a preset deviation (for example, the average minus 2 times the standard deviation), the frame is determined to be in the inefficient and high-consumption operating zone.

[0156] It should be noted that those skilled in the art should understand that the implementation functions of each module shown in the above-described embodiments of the adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback can be understood with reference to the relevant descriptions of the aforementioned adaptive control method for steel production equipment based on instantaneous energy efficiency characteristic feedback. The functions of each module shown in the above-described embodiments of the adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback can be implemented by a program (executable instructions) running on a processor, or by specific logic circuits. If the above-described adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part 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), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any particular combination of hardware and software.

[0157] Accordingly, embodiments of this application also provide a computer storage medium storing computer-executable instructions, which, when executed by a processor, implement the various method embodiments of this application.

[0158] Furthermore, this application also provides an adaptive control system for steel production equipment based on instantaneous energy efficiency characteristic feedback, including a memory for storing computer-executable instructions and a processor; the processor is used to implement the steps in the above-described method embodiments when executing the computer-executable instructions in the memory. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The aforementioned memory can be read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or solid-state drive, etc. The steps of the methods disclosed in the embodiments of this application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor.

[0159] It should be noted that in this patent application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this patent application, if it refers to performing an action according to an element, it means performing the action at least according to that element, including two cases: performing the action only according to that element, and performing the action according to that element and other elements. Expressions such as "multiple," "repeatedly," and "various" include two, two times, two kinds, and more than two, more than two times, and more than two kinds.

[0160] All documents mentioned in this application are considered to be incorporated in their entirety into the disclosure of this application so that they can serve as a basis for modifications if necessary. Furthermore, it should be understood that after reading the foregoing disclosure of this application, those skilled in the art can make various alterations or modifications to this application, and these equivalent forms also fall within the scope of protection claimed in this application.

Claims

1. An adaptive control method for steel production equipment based on instantaneous energy efficiency characteristic feedback, characterized in that, Includes the following steps: S1: Through sensor groups deployed in key process sections of the steel production line, equipment operating status data Ddev and material flow status data Dmat are synchronously collected at millisecond intervals; The device operating status data Ddev includes at least: the instantaneous flow rate of the fuel medium and the current value of the drive motor; The material flow status data Dmat includes at least: temperature distribution and geometric dimension data of the work-in-process obtained by infrared imaging or laser scanning; S2: Using the edge computing unit, the equipment operation status data Ddev and material flow status data Dmat collected in step S1 are input into a preset energy consumption-material coupling model, and the instantaneous characteristic value Einst(t) representing the energy consumption intensity per unit of effective output in the current sampling period is calculated according to the following formula: in, The power consumption of the equipment during the current sampling period is obtained by multiplying the instantaneous flow rate of the fuel medium by the preset or online detected calorific value of the fuel, and summing the power consumption of the drive motor. The effective material output rate within the current sampling period is calculated from the real-time cross-sectional area of ​​the work-in-process obtained by combining the roller conveyor weighing signal and speed measurement signal with laser diameter measurement; the unit of the instantaneous characteristic value Einst(t) is kJ / kg. When the material input remains constant but the effective output decreases, the instantaneous characteristic value Einst(t) shows an instantaneous pulse-like increase. S3: Compare the instantaneous feature value Einst(t) obtained in step S2 with the dynamic benchmark threshold to determine whether there is an energy efficiency deviation; When the instantaneous characteristic value Einst(t) is detected to exceed the upper limit of the dynamic reference threshold, the process stability constraint model is further invoked to calculate the equipment safety index Sstab in real time based on the data collected in step S1, and to verify whether the current equipment is within the allowable adjustment safety margin range. The process stability constraint model includes the following constraint indices: Furnace temperature uniformity variance constraint: Calculate the temperature variance based on the temperature data of multiple thermocouples in the furnace, and determine whether the temperature variance is less than the upper limit of variance set in the process specification. Plate shape straightness constraint: Based on the real-time feedback signal of the plate shape instrument, the wave shape index is obtained, and it is determined whether the wave shape index is within the preset controlled range; Motor temperature rise rate constraint: The temperature rise rate is calculated based on data from the motor winding temperature sensor, and it is determined whether the temperature rise rate is lower than the safety threshold. The physical correction instruction will only be issued to the actuator if the generated physical correction instruction is predicted by simulation to not cause the above-mentioned constraint indicators to exceed the limit. S4: When step S3 determines that the instantaneous feature value Einst(t) deviates abnormally and the equipment safety index Sstab meets the safety margin requirements, a physical correction instruction for the production equipment actuator is generated based on the control algorithm and sent to the process control system through the industrial communication interface. The physical correction command drives the actuator to adjust the physical operating parameters, so that the instantaneous feature value Einst(t) returns to the dynamic benchmark threshold range, forming a closed-loop feedback; Step S4 also includes a cross-process energy consumption dynamic compensation strategy for material volume loss: When the laser scanning data in step S1 indicates that the measured volume of the intermediate billet at the roughing mill exit is less than the theoretically set threshold, it indicates that excessive cutting of the head and tail or large oxidation and burning loss has caused irreversible volume loss of the material, which increases the instantaneous characteristic value Einst(t) calculated in step S2. To offset the increase in unit energy consumption caused by irreversible material loss in the preceding process, the physical correction command is applied to the subsequent laminar flow cooling process: under the condition of meeting the process deviation constraint of the winding temperature, the low-pressure section cooling manifold is automatically and preferentially turned on, and the inverter output frequency of the high-pressure water supply pump is reduced or turned off. By reducing the electrical energy consumption in the cooling process, material loss in the preceding processes can be physically compensated, thereby achieving energy efficiency optimization throughout the entire process.

2. The method according to claim 1, characterized in that, In step S1: The equipment operating status data Ddev also includes: the torque of the main motor of the rolling mill, the frequency of the combustion fan, the power of the cooling water pump, and the pressure of the descaling pump; The material flow status data Dmat also includes: instantaneous throughput based on roller conveyor weighing sensors and material conveying speed based on speed encoders.

3. The method according to claim 1, characterized in that, In step S3, the dynamic benchmark threshold is automatically generated based on historical stable operating condition data, and its generation method is as follows: Sliding window statistics are performed on the instantaneous characteristic value Einst(t) of qualified working conditions within the most recent preset time window, and the mean and standard deviation within the window are calculated. The upper limit of the threshold is set by adding a preset multiple of the standard deviation to the mean; When the equipment is detected to be in an unsteady state of switching steel grades or changing rolls, the update of the dynamic reference threshold is paused.

4. The method according to claim 1, characterized in that, In step S4, the actuator includes at least one of the following: The gas regulating valve of the heating furnace is used to adjust the valve opening to control the fuel supply. The frequency converter of the combustion blower is used to adjust the blower frequency to control the combustion air volume; The roll gap servo mechanism of the rolling mill is used to adjust the roll gap setpoint to control the reduction amount; The frequency converter of the cooling system's water pump is used to adjust the pump's operating frequency to control the cooling water flow rate.

5. The method according to claim 1, characterized in that, For the heating furnace area, the control strategy in step S4 includes: When step S1 detects that the speed of the furnace feed rollers decreases, causing the material flow to slow down, but the gas flow rate remains high, the instantaneous characteristic value Einst(t) calculated in step S2 increases, and it is determined to be a state of high consumption due to dry burning. In step S3, if the surface temperature of the steel billet in the furnace has reached the upper limit of the process and the center temperature meets the process requirements, it is determined that the stability constraint allows for cooling. The physical correction instructions include: reducing the opening of the gas regulating valve in the heat exchange section, and / or reducing the output frequency of the combustion fan inverter, thereby reducing excess air heat loss and reducing the supply of ineffective calorific value.

6. The method according to claim 1, characterized in that, For the rolling zone, the control strategy in step S4 includes: Based on the main motor current data of each rack collected in step S1, when an abnormal high-frequency fluctuation is detected in the main motor current of a rack, it is identified that the instantaneous energy efficiency ratio of the rack is lower than the preset lower limit threshold of energy efficiency ratio, and it is in the low-efficiency and high-consumption operation zone. Using an automatic thickness control system, while maintaining a constant total reduction rate, a portion of the reduction load of this frame is dynamically distributed to adjacent high-efficiency frames. This allows the entire unit to operate within the high-efficiency range of the motor, reducing overall power consumption.

Images