Fusion control method, device and equipment for hearth temperature of sintering ignition furnace

Abstract

According to the fusion control method, device and equipment for the hearth temperature of the sintering ignition furnace, by fusing the characteristic matrix of multi-source parameters, a hearth temperature prediction model can more comprehensively and accurately sense changes of the production process, accurate prediction of the hearth temperature is achieved, and therefore passive response is changed into active look-ahead control; through a data-driven hearth temperature prediction model dominant and physical static heat balance model verification bottom-filling hybrid architecture, the advantage of processing complex nonlinearity of the hearth temperature prediction model is played, the rationality and safety of control are ensured by using a physical model, the severe change of a working condition can be effectively handled, and the service life of the system is prolonged. The control instruction is prevented from violating a basic physical rule; by means of feedforward-feedback fusion control, stable and accurate adjustment of the temperature of the hearth is achieved, over-burning and under-burning are avoided through accurate temperature control, the quality of sintered ore is stabilized and improved, meanwhile, gas consumption is effectively reduced, and the requirement for green intelligent manufacturing is met.

Description

Technical Field

[0001] This invention relates to the field of sintering technology, and in particular to a method, apparatus and equipment for fusion control of furnace temperature in a sintering ignition furnace. Background Technology

[0002] Sintering is a crucial process in steel production, transforming powdered iron-containing raw materials into high-strength, high-metallurgical-performance sintered ore. The sintering ignition furnace, as the starting point and core equipment of the sintering process, has the primary task of providing stable, uniform, and suitable heat to the surface of the sintering material layer to ignite the surface fuel, laying the foundation for the subsequent automatic regenerative sintering process. The stability, accuracy, and matching degree of the furnace temperature with the material state directly determine the yield, quality (such as drum strength and FeO content) of the sintered ore, as well as energy consumption.

[0003] Currently, the temperature control of sintering ignition furnaces mainly follows these modes: (1) Manual control: This method relies entirely on the operator's experience to observe the instruments and manually adjust the opening of the gas and air valves. This method has a slow response, low control accuracy, poor stability, and cannot adapt to frequent changes in raw materials and working conditions, resulting in large fluctuations in product quality and high energy consumption.

[0004] (2) Traditional PID control: Establish a PID controller for furnace temperature and gas / air flow. However, the sintering process has the characteristics of large inertia, large lag and strong nonlinearity. The parameters of the traditional PID controller are fixed, making it difficult to make timely and appropriate adjustments to complex working conditions. It generally has the disadvantages of large overshoot, weak anti-interference ability and long settling time.

[0005] (3) Early intelligent control: Some advanced solutions attempt to introduce algorithms such as neural networks and fuzzy control. However, these methods often have limitations: First, the input features are not well considered, and the "slow variables" that determine heat demand, such as sintering machine speed, material layer thickness, and material moisture, are not deeply integrated with the "fast variables" of thermal performance; Second, the model update mechanism is lacking, and once the trained model is deployed, it remains fixed and cannot adapt to equipment aging and changes in raw material characteristics; Third, there is a lack of physical law constraints, and the purely data-driven model may output control commands that violate thermodynamic principles when the operating conditions change drastically, which poses a safety risk.

[0006] In summary, existing technologies have failed to effectively address the core issue of how to achieve high-precision, real-time, adaptive intelligent prediction and control of the furnace temperature under complex and variable sintering conditions.

[0007] In view of this, it is necessary to propose a method, device and equipment for the integrated control of the furnace temperature of a sintering ignition furnace in order to solve or at least alleviate the above-mentioned defects. Summary of the Invention

[0008] The main objective of this invention is to provide a method, apparatus, and equipment for integrated control of the furnace temperature of a sintering ignition furnace, in order to solve the problem of how to achieve high-precision, real-time, and adaptive control of the furnace temperature under complex and variable sintering conditions.

[0009] To achieve the above objectives, the present invention provides a method for fusion control of the furnace temperature of a sintering ignition furnace, comprising the following steps: S1, obtain the current real-time sintering condition parameters, input the real-time sintering condition parameters into the pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detection flow rate and the furnace detection temperature. S2, input the current real-time operating parameters and target furnace temperature into the pre-calibrated static heat balance model, and receive the theoretically reasonable gas flow rate output by the static heat balance model; S3, obtain the gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main; S4, the gas flow feedforward command is fused with the gas flow feedback command from the pre-built independent PID feedback controller to obtain the fused control command; S5, compare the fusion control command with the theoretically reasonable gas flow rate, and determine the final gas flow control command based on the comparison result; S6, based on the final gas flow control command and the preset air-fuel ratio, synchronously adjust the gas and air regulating valve actuators.

[0010] Preferably, the furnace temperature prediction model in step S1 is obtained through the following steps: S11, collect real-time sintering condition parameters within a preset time period to obtain an initial dataset, and preprocess the initial dataset to obtain a historical dataset for constructing a furnace temperature prediction model. S12, after randomly shuffling the historical dataset to eliminate time series bias, it is divided into mutually exclusive training set, validation set and test set according to the proportion; S13, The backpropagation neural network model is selected as the core prediction model, and a genetic algorithm is introduced to globally optimize the initial weights and biases of the backpropagation neural network model to obtain the GA-BP neural network model. S14, the GA-BP neural network model is trained using the training set and its performance is monitored on the validation set. The optimal training rounds are determined by the early stopping method to obtain the furnace temperature prediction model.

[0011] Preferably, the static thermal equilibrium model in step S2 is obtained through the following steps: S21. Based on the law of conservation of energy in the ignition furnace, the static thermal balance model is established. S22. During the period when the sintering machine is in stable operation, a set of real-time sintering condition parameters are collected, and the key empirical parameters of the static thermal balance model are determined based on the real-time sintering condition parameters to complete the verification of the static thermal balance model and obtain the verified static thermal balance model.

[0012] Preferably, the expression for the static thermal equilibrium model is: ;in, Gas flow rate, unit: m³ 3 / h, HCV is the calorific value of coal gas, in kJ / m³. 3 η is the overall combustion efficiency, and α is the effective heat absorption coefficient. The specific heat capacity of the material is expressed in kJ / (kg·℃). The temperature rise of the material is expressed in degrees Celsius (°C). This refers to the mass flow rate of the material, expressed in kg / h. The unit is system heat loss, expressed in kJ / h.

[0013] Preferably, the preprocessing of the initial dataset in step S11 includes the following steps: S111, perform data cleaning on the initial dataset; S112, perform time-series alignment on the initial dataset after cleaning; wherein, the sampling frequency of all parameters in the initial dataset is unified to a fixed time interval; based on the sintering machine speed, calculate the lag time required for the material to travel from the detection point to directly below the ignition furnace; based on the lag time, match the material parameters detected in advance with the thermal parameters of the ignition furnace after the lag time to perform time-series alignment. S113, normalize the dataset processed in step S112.

[0014] Preferably, step S3 specifically includes the following steps: Using formula Calculate the gas flow feedforward command ;in, The current flow rate of the main gas pipe is measured in meters (m³). 3 / h, The proportional gain of the feedforward controller, in meters. 3 / (h·℃), The target furnace temperature, in °C. This is the predicted furnace temperature, in °C.

[0015] Preferably, the step S5 of determining the final gas flow control command based on the comparison result includes the following steps: If the deviation between the fusion control command and the theoretically reasonable gas flow rate is within a preset safety boundary, then the fusion control command is used as the final gas flow control command. If the deviation exceeds the preset safety boundary, an alarm is triggered, and a final gas flow control command is generated using a conservative control strategy based on the theoretically reasonable gas flow rate. The conservative control strategy includes: directly using the theoretically reasonable gas flow rate, or using the weighted average of the fusion control command and the theoretically reasonable gas flow rate.

[0016] Preferably, the fusion control method further includes a model performance monitoring and updating step that is executed in parallel with the control process, specifically including the following steps: Obtain the actual furnace temperature and calculate the absolute value of the temperature difference between the actual furnace temperature and the predicted furnace temperature. When the absolute value of the temperature difference is greater than the first preset threshold, the gradient descent method is used to adjust the weights or biases of the furnace temperature prediction model in order to achieve slow adaptation of the furnace temperature prediction model. When the absolute value of the temperature difference is greater than the second preset threshold, it is determined that the current furnace temperature prediction model has failed, triggering the model reconstruction mechanism, returning to step S1, collecting new real-time sintering condition parameters, and retraining the furnace temperature prediction model; wherein, the second preset threshold is greater than the first preset threshold.

[0017] The present invention also provides a fusion control device for the furnace temperature of a sintering ignition furnace, used to execute the fusion control method for the furnace temperature of a sintering ignition furnace as described above, comprising: The temperature prediction module is used to acquire the current real-time sintering condition parameters, input the real-time sintering condition parameters into a pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detected flow rate and the furnace detected temperature. The physical verification module is used to input the current real-time operating parameters and target furnace temperature into the pre-verified static heat balance model, and to receive the theoretically reasonable gas flow rate output by the static heat balance model. The feedforward control module is used to obtain a gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main. The instruction fusion module is used to fuse the gas flow feedforward instruction with the gas flow feedback instruction from the pre-built independent PID feedback controller to obtain the fused control instruction. The safety decision module compares the fused control command with the theoretically reasonable gas flow rate and determines the final gas flow control command based on the comparison result. The execution control module is used to synchronously adjust the gas and air regulating valves according to the final gas flow control command and the preset air-fuel ratio.

[0018] The present invention also provides a fusion control device for the furnace temperature of a sintering ignition furnace, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the fusion control method for the furnace temperature of a sintering ignition furnace as described above.

[0019] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a fusion control method, device, and equipment for the furnace temperature of a sintering ignition furnace, achieving high-precision, real-time, and adaptive control of the furnace temperature under complex and variable sintering conditions. By fusing the feature matrix of multi-source parameters, the furnace temperature prediction model can more comprehensively and accurately perceive changes in the production process, achieving precise prediction of the furnace temperature, thus transforming passive response into proactive look-ahead control. Through a hybrid architecture of a data-driven furnace temperature prediction model as the primary driver and a physical static thermal balance model as a fallback, the invention leverages the advantages of the furnace temperature prediction model in handling complex nonlinearities while utilizing a physical model to ensure the rationality and safety of the control. This effectively copes with drastic changes in operating conditions and prevents control commands from violating basic physical laws. Through feedforward-feedback fusion control, stable and precise regulation of the furnace temperature is achieved. Precise temperature control avoids over-burning and under-burning, effectively reducing gas consumption while stabilizing and improving the quality of sintered ore, meeting the requirements of green and intelligent manufacturing. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0021] Figure 1 This is a flow chart of the sintering process in one embodiment of the present invention; Figure 2 This is a flowchart illustrating an embodiment of the present invention.

[0022] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0026] Furthermore, the use of terms such as "first" and "second" in this invention is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. If the combination of technical solutions is contradictory or impossible to implement, such a combination should be considered non-existent and not within the scope of protection claimed by this invention.

[0027] Please see the appendix Figures 1 to 2 The present invention provides a method for controlling the furnace temperature of a sintering ignition furnace according to an embodiment, comprising the following steps: S1, obtain the current real-time sintering condition parameters, input the real-time sintering condition parameters into the pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detection flow rate and the furnace detection temperature. Preferably, the thermal parameters of the ignition furnace include furnace temperature (°C), gas main pipe flow rate (m³ / h), gas main pipe pressure (kPa), air main pipe flow rate (m³ / h), air main pipe pressure (kPa), furnace negative pressure (Pa), and negative pressure of wind box #1 (kPa, wind box #2) (kPa). The sintering machine operating parameters include sintering machine speed (m / min); the sintering material parameters include material layer thickness (mm), sintering material moisture content (%), and pressing density (t / m²).

[0028] As an example, the following real-time sintering parameters were collected: sintering machine speed 2.15 m / min, material layer thickness 875 mm, furnace negative pressure -25 Pa, furnace temperature target 1100℃, #1 wind box negative pressure -15.5 kPa, #2 wind box negative pressure -15.6 kPa. Gas flow rate range 8000-15000 m³ / h, air flow rate range 6000-13000 m³ / h, and sintering material moisture content 7.0%.

[0029] S2, input the current real-time operating parameters and target furnace temperature into the pre-calibrated static heat balance model, and receive the theoretically reasonable gas flow rate output by the static heat balance model; wherein, the target furnace temperature can be preset and is a known quantity.

[0030] S3, obtain the gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main; Furthermore, step S3 specifically includes the step of: using the formula Calculate the gas flow feedforward command Unit m 3 / h; where, The current flow rate of the main gas pipe is measured in meters (m³). 3 / h can be used for detection and acquisition. The proportional gain of the feedforward controller, in meters. 3 / (h·℃), adjustments were made at the engineering site, for example, by using historical steady-state samples. A linear regression was performed between the flow rate and the main gas pipe flow rate, and the slope obtained from the fitted data was used as the proportional coefficient of the feedforward controller. The target furnace temperature, in °C. This is the predicted furnace temperature, in °C.

[0031] Specifically, the target furnace temperature and the predicted furnace temperature are input into a proportional feedforward controller to calculate the initial gas flow feedforward command. This proportional feedforward controller is used to eliminate the deviation between the target furnace temperature and the predicted furnace temperature, and to quickly compensate for disturbances in key, measurable operating parameters. If the predicted furnace temperature is lower than the target furnace temperature, the initial gas flow feedforward command will increase the current flow rate; conversely, it will decrease it. This step, through the furnace temperature prediction model combined with feedforward control, can proactively compensate for disturbances caused by material parameters, effectively overcoming the large inertia and hysteresis characteristics of the sintering process, reducing temperature control overshoot, and shortening the settling time.

[0032] S4, the gas flow feedforward command is fused with the gas flow feedback command from the pre-built independent PID feedback controller to obtain the fused control command; Specifically, an independent PID feedback controller can be obtained through the following steps: S41, Feedback signal acquisition: Real-time acquisition of the actual furnace temperature measured by thermocouples; S42, Deviation Calculation: Calculate the real-time deviation e(t) between the target furnace temperature and the actual furnace temperature; S43, PID Calculation: The deviation e(t) is input into the PID controller, and its output is the gas flow feedback command. The control law is: ;in, , , These are the proportional gain coefficients, in meters. 3 / (h·℃), integral coefficient, unit m 3 / (h·℃·s), differential coefficient, unit m 3 ·s / (h·℃) is the pre-tuned controller parameter. This is a gas flow feedback command, in meters (m). 3 / h, the core function of this controller is to eliminate any steady-state error caused by disturbances, ensuring that the system can eventually stabilize at the target furnace temperature without steady-state error.

[0033] Feedforward command for gas flow Gas flow feedback instructions The commands are added together to obtain the fused control command. This step ensures the steady-state accuracy of the control by introducing an independent PID feedback controller and fusing it with the feedforward command.

[0034] S5, compare the fusion control command with the theoretically reasonable gas flow rate, and determine the final gas flow control command based on the comparison result; S6, based on the final gas flow control command and the preset air-fuel ratio, synchronously adjust the gas and air regulating valve actuators.

[0035] In this application, by integrating the feature matrix of multi-source parameters, the furnace temperature prediction model can more comprehensively and accurately perceive changes in the production process, achieving precise prediction of furnace temperature and thus transforming passive response into proactive forward control. Through a hybrid architecture of data-driven furnace temperature prediction model as the primary driver and physical static thermal balance model as a fallback, the model leverages its advantage in handling complex nonlinearities while utilizing a physical model to ensure the rationality and safety of control. This effectively addresses drastic changes in operating conditions and prevents control commands from violating fundamental physical laws. Through feedforward-feedback fusion control, stable and precise regulation of the furnace temperature is achieved. Precise temperature control avoids over-burning and under-burning, effectively reducing gas consumption while stabilizing and improving sinter quality, thus meeting the requirements of green and intelligent manufacturing.

[0036] As a preferred embodiment, the furnace temperature prediction model in step S1 is obtained through the following steps: S11, collect real-time sintering condition parameters within a preset time period to obtain an initial dataset, and preprocess the initial dataset to obtain a historical dataset for constructing a furnace temperature prediction model. S12, after randomly shuffling the historical dataset to eliminate temporal bias, it is divided proportionally into mutually exclusive training, validation, and test sets. As a preferred example, the training set (70%–80% of the total) is used for model training and adjusting the model's internal parameters. The validation set (10%–15% of the total) is used to monitor model performance during training, optimize hyperparameters, and implement early stopping strategies to prevent overfitting. The test set (10%–15% of the total) is used for unbiased evaluation of the model's final generalization performance after the model training and tuning process is complete.

[0037] S13, The error backpropagation neural network model is selected as the core prediction model, and the genetic algorithm is introduced to globally optimize the initial weights and biases of the error backpropagation neural network model to obtain the GA-BP neural network model. Specifically, the sintering process is a complex nonlinear system, and the backpropagation neural network model (BP neural network model) has a strong nonlinear mapping capability and adaptive learning capability, which can accurately fit the complex mapping relationship between furnace temperature and various operating parameters.

[0038] It is worth noting that, to address the issues of standard BP neural network models easily getting trapped in local optima and exhibiting slow convergence, this invention introduces a genetic algorithm (GA) to globally optimize the initial weights and biases of the BP neural network model. GA simulates natural evolution and can provide a set of initial network parameters that approximate the global optimum, thereby improving model performance and stability. This combined model is called a GA-BP neural network, resulting in the GA-BP neural network model.

[0039] Network Structure: The GA-BP neural network model adopts a three-layer feedforward structure, denoted as IHO. Here, I represents the number of input layer nodes, equal to the number of feature parameters; H represents the number of hidden layer nodes, which can be determined through empirical formulas and experiments; and O represents the number of output layer nodes, which is 1 in this invention, representing the predicted furnace temperature.

[0040] Specifically, all the weights and biases of the BP neural network model can be arranged into a real number vector (chromosome). The MSE of the training set can be used as the fitness. The open library can be called to iterate 30-50 times to select the vector corresponding to the smallest MSE. Then, the optimal vector is decomposed back into the weight matrix and the standard BP neural network model is trained to obtain the GA-BP neural network model.

[0041] S14, the GA-BP neural network model is trained using the training set and its performance is monitored on the validation set. The optimal training rounds are determined using the early stopping method to obtain the furnace temperature prediction model. Further, when the performance indicators (such as mean square error MSE) of the furnace temperature prediction model on the test set reach a preset value, the furnace temperature prediction model can be deployed on the field controller.

[0042] As a preferred embodiment, the static thermal equilibrium model in step S2 is obtained through the following steps: S21. Based on the law of conservation of energy in the ignition furnace, the static thermal balance model is established. S22. During the period when the sintering machine is in stable operation, a set of real-time sintering condition parameters are collected, and the key empirical parameters of the static thermal balance model are determined based on the real-time sintering condition parameters to complete the verification of the static thermal balance model and obtain the verified static thermal balance model.

[0043] Furthermore, the expression for the static thermal equilibrium model is: ;in, Gas flow rate, unit: m³ 3 / h, HCV is the calorific value of coal gas, in kJ / m³. 3 η is the overall combustion efficiency, and α is the effective heat absorption coefficient. The specific heat capacity of the material is expressed in kJ / (kg·℃). The temperature rise of the material is expressed in degrees Celsius (°C). This refers to the mass flow rate of the material, expressed in kg / h. The unit is system heat loss, expressed in kJ / h.

[0044] This embodiment provides a theoretical benchmark for the entire control system based on the law of conservation of energy, ensuring the physical rationality of the core control logic. It also verifies key empirical parameters by collecting data under stable operating conditions and using this data to back-calculate the results. , , The values ​​of these parameters are used to complete the static thermal balance model verification, improving accuracy and relevance. This is because the parameters are taken from real-time sintering conditions during a period of stable operation. HCV , , Since all quantities are known, the least squares method can be used to regress them. , , (i.e., key empirical parameters) In real-time control, the static heat balance model can calculate the theoretically reasonable gas flow rate by substituting the current material parameters (such as bed thickness, machine speed, raw material moisture content, air-fuel ratio, etc.) and the target furnace temperature into the above formula.

[0045] In a preferred embodiment, the preprocessing of the initial dataset in step S11 includes the following steps: S111, perform data cleaning on the initial dataset; Specifically, linear interpolation is used to fill in short-term missing continuous parameters (such as flow rate and pressure); for long-term missing or critical but slowly updated material parameters (such as compressive density), the entire data sample corresponding to that time period is directly removed to avoid introducing erroneous information.

[0046] Outlier handling: A two-stage filtering process is employed. First, physical threshold filtering: reasonable upper and lower limits are set for each parameter based on process knowledge (e.g., the effective range of furnace temperature is 800℃~1200℃), eliminating obviously unreasonable data. Second, statistical filtering: the 3-Sigma criterion (Laida criterion) is used to identify and eliminate abnormal fluctuations hidden within a reasonable range. This criterion states that for data that approximately follows a normal distribution, 99.73% of the values ​​will fall within the interval [μ-3σ, μ+3σ] of the mean (μ) plus or minus three standard deviations (σ). Data falling outside this interval are considered low-probability outliers and are eliminated. μ and σ are calculated based on historical data under stable operating conditions. This step effectively removes noise and errors from the original industrial data.

[0047] S112, perform time-series alignment on the cleaned initial dataset; wherein, the sampling frequency of all parameters in the initial dataset is unified to a fixed time interval, such as 10s; for example, downsampling is performed on high-frequency parameters (such as temperature) (taking the average value within the period), and forward padding is performed on low-frequency parameters (such as moisture content) (keeping the previous valid value until the next update).

[0048] Based on the sintering machine speed, under stable operating conditions, the sintering machine can be considered to be moving at a constant speed. The lag time required for the material to travel from the detection point to directly below the ignition furnace is calculated. Based on the lag time, the material parameters detected in advance are matched with the thermal parameters of the ignition furnace after the lag time to perform time-series alignment. This constitutes an effective sample with causal relationship, ensuring that the model learns the thermal results produced by the material in a certain state in the ignition furnace.

[0049] As an example, it was calculated that it takes about 3.5 minutes for the material to travel from the material layer thickness detection point to the ignition furnace, and data matching was performed based on this.

[0050] S113, normalize the dataset processed in step S112. As a better example, Z-Score normalization is used to eliminate differences in units and orders of magnitude between different parameters, accelerating model convergence. Z-Score normalization is insensitive to outliers, and the processed data conforms to a standard normal distribution, which is more conducive to gradient descent algorithm optimization.

[0051] As an example: The average gas flow rate calculated from the training set is μ = 11000 m³ / h, σ = 1500 m³ / h. The real-time flow rate of 12500 m³ / h is normalized to (12500-11000) / 1500≈1.0.

[0052] In a preferred embodiment, step S5, which involves determining the final gas flow control command based on the comparison results, includes the following steps: If the deviation between the fusion control command and the theoretically reasonable gas flow rate is within a preset safety boundary, then the fusion control command is used as the final gas flow control command. If the deviation exceeds the preset safety boundary, an alarm is triggered, and a final gas flow control command is generated using a conservative control strategy based on the theoretically reasonable gas flow rate. The conservative control strategy includes: directly using the theoretically reasonable gas flow rate, or using the weighted average of the fusion control command and the theoretically reasonable gas flow rate.

[0053] Specifically, the fusion control command is compared with the theoretically reasonable gas flow rate, and a decision can be made according to the following rules: Normal mode: If the deviation between the fusion control command and the theoretically reasonable gas flow rate is within the preset safety boundary (e.g., ±10%), then the fusion control command that integrates feedforward and feedback is preferred as the final gas flow control command.

[0054] Safety mode: If the deviation exceeds the safety boundary, an alarm will be triggered and a conservative strategy will be adopted: either the output of the static thermal balance model will be based on the theoretically reasonable gas flow rate, or a weighted average of the two will be performed (such as giving the static thermal balance model a higher weight) to ensure the physical rationality of the control command. This option is manually selected and confirmed, and the generated conservative command will be used as the final gas flow control command.

[0055] As a preferred example, the theoretically reasonable gas flow rate can be directly used, or a weighted average of the fusion control command and the theoretically reasonable gas flow rate can be used, with the theoretically reasonable gas flow rate given a higher weight (e.g., fusion control command: 0.3, theoretically reasonable gas flow rate: 0.7). For example, if the fusion control command is 14000 m³ / h and the theoretically reasonable gas flow rate is 12500 m³ / h, and the deviation between the two is 12%, then the safety mode is entered.

[0056] This embodiment sets a preset safety boundary and prioritizes the use of fused control commands within this boundary. When the boundary is exceeded, a conservative control strategy based on the theoretically reasonable gas flow rate obtained from the static thermal balance model is switched on. This effectively prevents the purely data-driven model from outputting dangerous commands that violate physical laws under abnormal operating conditions, thus avoiding equipment damage and production accidents. An alarm mechanism is triggered to promptly notify operators and allow for timely manual intervention.

[0057] In another preferred embodiment, the fusion control method further includes a model performance monitoring and updating step that is executed in parallel with the control process, specifically including the following steps: Obtain the actual furnace temperature and calculate the absolute value of the temperature difference between the actual furnace temperature and the predicted furnace temperature. When the absolute value of the temperature difference is greater than the first preset threshold, the gradient descent method is used to adjust the weights or biases of the furnace temperature prediction model in order to achieve slow adaptation of the furnace temperature prediction model. When the absolute value of the temperature difference is greater than the second preset threshold, it is determined that the current furnace temperature prediction model has failed, triggering the model reconstruction mechanism, returning to step S1, collecting new real-time sintering condition parameters, and retraining the furnace temperature prediction model; wherein, the second preset threshold is greater than the first preset threshold.

[0058] Specifically, by continuously monitoring the absolute value of the temperature difference between the furnace temperature prediction value output by the furnace temperature prediction model and the actual furnace temperature, when the absolute value of the temperature difference is greater than a first preset threshold, an online learning algorithm (such as gradient descent) can be used to fine-tune the connection weights and bias parameters of the deployed furnace temperature prediction model, so as to achieve slow self-adaptation of the model. This solves the problem that traditional models are fixed once deployed, and this embodiment can be updated online adaptively, which greatly reduces the need for manual maintenance.

[0059] When the absolute value of the temperature difference is greater than the second preset threshold, or when the absolute value of the temperature difference continues to exceed a large preset threshold, or when a significant step change in production conditions is detected (such as a large change in the calorific value of the gas or a switch of the main raw materials), the current furnace temperature prediction model is determined to be invalid. The system automatically triggers the model reconstruction mechanism, returns to step S1, collects new field data, restarts the model training process, and ensures the long-term effectiveness of the model.

[0060] As a preferred example, a first preset threshold can be set to a deviation of 2%, and a second preset threshold can be set to a deviation of 5%. For example, when the calorific value of the gas changes abruptly from 3800 kJ / m³ to 3600 kJ / m³, the system automatically triggers model reconstruction.

[0061] The present invention also provides a fusion control device for the furnace temperature of a sintering ignition furnace, used to execute the fusion control method for the furnace temperature of a sintering ignition furnace as described above, comprising: The temperature prediction module is used to acquire the current real-time sintering condition parameters, input the real-time sintering condition parameters into a pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detected flow rate and the furnace detected temperature. The physical verification module is used to input the current real-time operating parameters and target furnace temperature into the pre-verified static heat balance model, and to receive the theoretically reasonable gas flow rate output by the static heat balance model. The feedforward control module is used to obtain a gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main. The instruction fusion module is used to fuse the gas flow feedforward instruction with the gas flow feedback instruction from the pre-built independent PID feedback controller to obtain the fused control instruction. Safety decision module: compares the fused control command with the theoretically reasonable gas flow rate, and determines the final gas flow control command based on the comparison result; The execution control module is used to synchronously adjust the gas and air regulating valves according to the final gas flow control command and the preset air-fuel ratio.

[0062] The present invention also provides a fusion control device for the furnace temperature of a sintering ignition furnace, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the fusion control method for the furnace temperature of a sintering ignition furnace as described above.

[0063] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for controlling the temperature of a sintering furnace chamber, characterized in that, Includes the following steps: S1, obtain the current real-time sintering condition parameters, input the real-time sintering condition parameters into the pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detection flow rate and the furnace detection temperature. S2, input the current real-time operating parameters and target furnace temperature into the pre-calibrated static heat balance model, and receive the theoretically reasonable gas flow rate output by the static heat balance model; S3, obtain the gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main; S4, the gas flow feedforward command is fused with the gas flow feedback command from the pre-built independent PID feedback controller to obtain the fused control command; S5, compare the fusion control command with the theoretically reasonable gas flow rate, and determine the final gas flow control command based on the comparison result; S6, based on the final gas flow control command and the preset air-fuel ratio, synchronously adjust the gas and air regulating valve actuators.

2. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 1, characterized in that, The furnace temperature prediction model in step S1 is obtained through the following steps: S11, collect real-time sintering condition parameters within a preset time period to obtain an initial dataset, and preprocess the initial dataset to obtain a historical dataset for constructing a furnace temperature prediction model. S12, after randomly shuffling the historical dataset to eliminate time series bias, it is divided into mutually exclusive training set, validation set and test set according to the proportion; S13, The backpropagation neural network model is selected as the core prediction model, and a genetic algorithm is introduced to globally optimize the initial weights and biases of the backpropagation neural network model to obtain the GA-BP neural network model. S14, the GA-BP neural network model is trained using the training set and its performance is monitored on the validation set. The optimal training rounds are determined by the early stopping method to obtain the furnace temperature prediction model.

3. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 1, characterized in that, The static thermal equilibrium model in step S2 is obtained through the following steps: S21. Based on the law of conservation of energy in the ignition furnace, the static thermal balance model is established. S22. During the period when the sintering machine is in stable operation, a set of real-time sintering condition parameters are collected, and the key empirical parameters of the static thermal balance model are determined based on the real-time sintering condition parameters to complete the verification of the static thermal balance model and obtain the verified static thermal balance model.

4. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 3, characterized in that, The expression for the static thermal equilibrium model is: ;in, η is the gas flow rate, HCV is the gas calorific value, η is the overall combustion efficiency, and α is the effective heat absorption coefficient. The specific heat capacity of the material. For material temperature rise, For material mass flow rate, This represents the system's heat loss.

5. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 2, characterized in that, The preprocessing of the initial dataset in step S11 includes the following steps: S111, perform data cleaning on the initial dataset; S112, perform time-series alignment on the initial dataset after cleaning; wherein, the sampling frequency of all parameters in the initial dataset is unified to a fixed time interval; based on the sintering machine speed, calculate the lag time required for the material to travel from the detection point to directly below the ignition furnace; based on the lag time, match the material parameters detected in advance with the thermal parameters of the ignition furnace after the lag time to perform time-series alignment. S113, normalize the dataset processed in step S112.

6. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 1, characterized in that, Step S3 specifically includes the following steps: Using formula Calculate the gas flow feedforward command ;in, The current gas main pipe is being monitored for flow rate. This is the proportional coefficient of the feedforward controller. For the target furnace temperature, This is the predicted value for furnace temperature.

7. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 1, characterized in that, The step S5, which determines the final gas flow control command based on the comparison results, includes the following steps: If the deviation between the fusion control command and the theoretically reasonable gas flow rate is within a preset safety boundary, then the fusion control command is used as the final gas flow control command. If the deviation exceeds the preset safety boundary, an alarm is triggered, and a final gas flow control command is generated using a conservative control strategy based on the theoretically reasonable gas flow rate. The conservative control strategy includes: directly using the theoretically reasonable gas flow rate, or using the weighted average of the fusion control command and the theoretically reasonable gas flow rate.

8. The method for controlling the furnace temperature of a sintering ignition furnace according to claim 1, characterized in that, The fusion control method also includes a model performance monitoring and updating step that is executed in parallel with the control process, specifically including the following steps: Obtain the actual furnace temperature and calculate the absolute value of the temperature difference between the actual furnace temperature and the predicted furnace temperature. When the absolute value of the temperature difference is greater than the first preset threshold, the gradient descent method is used to adjust the weights or biases of the furnace temperature prediction model in order to achieve slow adaptation of the furnace temperature prediction model. When the absolute value of the temperature difference is greater than the second preset threshold, it is determined that the current furnace temperature prediction model has failed, triggering the model reconstruction mechanism, returning to step S1, collecting new real-time sintering condition parameters, and retraining the furnace temperature prediction model; wherein, the second preset threshold is greater than the first preset threshold.

9. A fusion control device for the furnace temperature of a sintering ignition furnace, used to execute the fusion control method for the furnace temperature of a sintering ignition furnace according to any one of claims 1-8, characterized in that, include: The temperature prediction module is used to acquire the current real-time sintering condition parameters, input the real-time sintering condition parameters into a pre-built furnace temperature prediction model, and receive the furnace temperature prediction value output by the furnace temperature prediction model; wherein, the real-time sintering condition parameters include ignition furnace thermal state parameters, sintering machine operating state parameters, and sintering material parameters, and the ignition furnace thermal state parameters include at least the gas main pipe detected flow rate and the furnace detected temperature. The physical verification module is used to input the current real-time operating parameters and target furnace temperature into the pre-verified static heat balance model, and to receive the theoretically reasonable gas flow rate output by the static heat balance model. The feedforward control module is used to obtain a gas flow feedforward command based on the target furnace temperature, the predicted furnace temperature, and the detected flow rate of the gas main. The instruction fusion module is used to fuse the gas flow feedforward instruction with the gas flow feedback instruction from the pre-built independent PID feedback controller to obtain the fused control instruction. The safety decision module compares the fused control command with the theoretically reasonable gas flow rate and determines the final gas flow control command based on the comparison result. The execution control module is used to synchronously adjust the gas and air regulating valves according to the final gas flow control command and the preset air-fuel ratio.

10. A fusion control device for the furnace temperature of a sintering ignition furnace, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the fusion control method for the furnace temperature of a sintering ignition furnace as described in any one of claims 1 to 8.

Images