Roller mill combustion chamber outlet temperature control method and system

Abstract

The invention provides a roller mill combustion chamber outlet temperature control method and system, and the method comprises the steps: S1, building a valve opening-gas flow relation model of a roller mill combustion chamber, carrying out the recognition of the valve opening-gas flow relation model, and obtaining the parameters of the valve opening-gas flow relation model; s2, establishing a gas flow-outlet temperature relation model of the roller mill combustion chamber, and identifying the gas flow-outlet temperature relation model by using a regression identification algorithm based on adaptive gradient to obtain parameters of the gas flow-outlet temperature relation model; and S3, according to the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model, the outlet temperature is stably controlled through an algorithm of automatically adjusting the valve opening. The temperature of the outlet of the combustion chamber of the roller mill can be stably controlled.

Description

Technical Field

[0001] This invention relates to the field of blast furnace ironmaking technology, and in particular to a method and system for controlling the outlet temperature of a roller mill combustion chamber. Background Technology

[0002] Iron pellets are a key raw material for blast furnace ironmaking, and their metallurgical properties directly affect the permeability of the furnace charge, the reduction rate, and energy consumption. Among various pellet production methods, the vertical shaft furnace roasting process is widely used in modern pellet plants due to its advantages such as low energy consumption, high output, and small footprint. The vertical shaft furnace roasting process includes multiple stages such as grinding, pelletizing, drying, preheating, roasting, and cooling. Especially in the grinding stage, the thermal conditions directly affect the uniformity of pellet particle size and subsequent roasting efficiency, thus becoming a critical link in ensuring the quality and energy efficiency of the finished product.

[0003] In roller mills, the combustion chamber outlet temperature is a core control variable during the grinding stage, and its stability directly determines the strength and metallurgical properties of the finished pellets. Typically, the combustion gas serves as the primary heat source for the combustion chamber, and the outlet temperature is indirectly controlled by adjusting the gas flow rate through valve opening. However, this control process exhibits significant nonlinear and time-varying characteristics, influenced by valve flow characteristics, the dynamic stability of the gas supply system, and variations in gas pressure. This presents challenges in determining the relationship between valve opening and gas flow rate, the relationship between gas flow rate and outlet temperature, and the control of the outlet temperature. Summary of the Invention

[0004] To address the technical problems existing in the prior art, the present invention provides a method and system for controlling the outlet temperature of a roller mill combustion chamber, the technical solution of which is as follows: On the one hand, a method for controlling the outlet temperature of the combustion chamber of a roller mill is provided, the method comprising: S1. Establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model; S2. Establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill, and use an adaptive gradient-based regression identification algorithm to identify the gas flow rate-outlet temperature relationship model to obtain the parameters of the gas flow rate-outlet temperature relationship model. S3. Based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model, the outlet temperature is stably controlled by an autonomous valve opening adjustment algorithm.

[0005] On the other hand, a combustion chamber outlet temperature control system for a roller mill is provided, the system comprising: The first identification module is used to establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model. The second identification module is used to establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill. The model is identified using an adaptive gradient-based regression identification algorithm to obtain the parameters of the gas flow rate-outlet temperature relationship model. The stability control module is used to stabilize the outlet temperature by autonomously adjusting the valve opening algorithm based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model.

[0006] On the other hand, an electronic device is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the above-described method for controlling the outlet temperature of the combustion chamber of a roller mill.

[0007] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, the at least one instruction being loaded and executed by a processor to implement the above-described method for controlling the outlet temperature of the combustion chamber of a roller mill.

[0008] The beneficial effects of the technical solution provided by this invention include at least the following: This invention considers the time-varying parameters of the gas flow-outlet temperature model and designs a regression identification based on adaptive gradients. This not only achieves consistent identification but also avoids drastic fluctuations in parameter estimation caused by noise points. Furthermore, since the valve opening adjustment is a fixed value each time, a composite control strategy based on an accumulator is proposed to prevent the outlet temperature from deviating from the expected range due to accumulated identification errors and measurement noise. This enables reliable estimation and control of the outlet temperature under information-constrained conditions, ensuring the normal operation of the roller mill. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of a method for controlling the outlet temperature of a roller mill combustion chamber according to an embodiment of the present invention; Figure 2This is a system block diagram of an embodiment of the present invention for determining the relationship between the valve opening degree and the gas flow rate in the combustion chamber of a roller mill; Figure 3 This is a block diagram of a combustion chamber outlet temperature control system for a roller mill provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0011] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0012] This invention provides a method for controlling the outlet temperature of a roller mill combustion chamber. This method can be implemented by an electronic device, which can be a terminal or a server. Figure 1 The diagram shown is a flowchart of the method. The processing flow may include the following steps:

[0013] S1. Establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model; Optionally, S1 specifically includes: exist At that moment, the valve opening degree was So when the gas pressure In the interval , At that time, the following valve opening-gas flow relationship model is established: (1) In model (1): For valve adjustable ratio, The maximum valve opening is dimensionless. The valve opening degree-gas flow rate relationship model is identified to obtain the parameters of the valve opening degree-gas flow rate relationship model.

[0014] More specifically, S1 includes: S11. Divide the fluctuation range of the combustion gas pressure in the roller mill combustion chamber into multiple intervals, and establish a valve opening-gas flow relationship model in each interval; S12. The valve opening-gas flow rate relationship model is transformed into a linear model in which the logarithm of the gas flow rate and the valve opening show a linear relationship. S13. Establish a measurement value model to describe the saturation measurement characteristics of the sensor that observes the gas flow rate, and establish a measurement label vector for the gas flow rate measurement value. S14. Count the frequency of each dimension of the measurement marker vector being 1; S15. Based on the frequency, the parameters of the linear model are identified using an empirical measure identification algorithm (empirical measure identification is frequency-based identification) to obtain estimated values ​​of the parameters of the linear model. S16. Based on the estimated values ​​of the parameters of the linear model, calculate the parameters of the valve opening-gas flow relationship model, and determine the relationship between the valve opening and gas flow in the combustion chamber of the roller mill based on the parameters of the valve opening-gas flow relationship model.

[0015] Optionally, S11 specifically includes: Under different gas pressures, the same valve opening will produce different gas flow rates, thus affecting the gas pressure. The fluctuation range is divided into A range, in the range , Internally, the relationship between valve opening and gas flow rate is considered constant, and the maximum flow rate that the gas valve can control is... and minimum flow Assuming a constant flow rate, the combustion chamber uses butterfly valves, which ensures a consistent gas flow rate. With valve opening The flow characteristics are equal percentages between them. At that moment, the valve opening degree was So when the gas pressure In the interval At that time, the following valve opening-gas flow relationship model is established: (1) In model (1): For valve adjustable ratio, The maximum valve opening is dimensionless.

[0016] Valve opening With actual gas flow The relationship between the valve opening and the gas flow rate exhibits strong nonlinearity and time-varying characteristics, which increases the difficulty of identifying the model parameters. Therefore, in this embodiment of the invention, the valve opening-gas flow rate relationship model is first transformed into a linear model in which the logarithm of the gas flow rate and the valve opening exhibit a linear relationship.

[0017] Optionally, S12 specifically includes: Taking the logarithm of both sides of model (1) and rearranging, we get: (2) In the formula: For the parameters to be identified, Treat it as model input; In model (2), the logarithm of the gas flow rate With valve opening There is a linear relationship between them.

[0018] Optionally, the measurement model established in S13 to describe the saturation measurement characteristics of the sensor observing the gas flow rate specifically includes: The sensor used to observe gas flow has a saturation measurement characteristic: when the actual gas flow exceeds the sensor's range... When the sensor output is at the upper limit of its range, the following measurement value model is established to describe this characteristic: (3) In model (3): It is the sensor in Real-time traffic The measured value, It is additional unknown noise, whose distribution function is expressed as: The unit is kPa.

[0019] The saturation measurement characteristic means that when the gas flow exceeds the sensor's range, the sensor cannot accurately output the true value, resulting in a loss of the ability to distinguish high flow conditions, which in turn negatively affects the identification effect of the valve opening-gas flow model parameters.

[0020] Optionally, in S13, the measured value of the gas flow rate is used to establish a measurement flag vector, specifically including: Because the sensor has saturation measurement characteristics, the measured value The relationship between the valve opening and the valve opening is no longer linear, nor is it simply described by model (2). Instead, it needs to be jointly represented with model (3) using measured values. With valve opening When identifying the valve opening-gas flow relationship model, it is necessary to avoid the parameter identification results from non-convergence caused by frequent transitions between linear and nonlinear relationships. Therefore, the data type of the input data should be standardized, and the measured values ​​during unsaturated measurements should be standardized. The quantification process is performed as follows: Select common There are several distinct thresholds, and there are ; Measurement values ​​at each time point Establish a set of measurement marker vectors ,in (4).

[0021] Optionally, S14 specifically includes: To ensure the health of the equipment, the valve opening is fixed each time it is adjusted. ( (usually a small percentage), all valve openings are recorded as... And there are ,in, It is the minimum valve opening, input vector ,in, ; In the pressure range Within this context, not all valve openings will occur, therefore, assuming for... , ,in, Indicates within the pressure range In the middle, the valve opening degree is The total number of time points, for , Therefore, for any pressure range From the front input vectors The regression matrix formed This is called the persistent feature matrix. (The last part is a separate, unrelated statement.) It is full rank, that is When the rank is 2, it is called the pressure zone. The corresponding model is called identifiable; When the measurement data given by the sensor is obtained At that time, firstly Quantization is performed to obtain the corresponding measurement label vector. ; Then, determine the pressure at the current moment. The range in which it is located, when the pressure In the interval And the valve opening degree is At that time, the actual measurement marker vector was obtained. The The frequency of each dimension being 1 is .

[0022] Optionally, S15 specifically includes: Identification is performed using the following formula: (5) (6) (7) In the formula: Pressure range Estimates of the internal model parameters, These are the local parameter estimates obtained based on each dimension of the measurement label vector. These are all weight parameters greater than 0, used to adjust the weights of local estimates, and for any... All have , It's noise. The inverse function of the cumulative distribution function, It is an indicator function, that is, when the event... At the time of its establishment, ,on the contrary, .

[0023] Optionally, the measurement marker vector The The probability that each dimension is 1 is: (8) when When, according to the law of large numbers: ; Therefore, we can conclude that: as well as (9), That is, parameter estimates It can converge to the truth value. .

[0024] Optionally, S16 specifically includes: Obtain parameter estimates Then, combined with what is known ,pass Solving for the given information yields the following results. , ; Furthermore, based on Solving for the given information yields the following results. ; Will Substituting into model (1), the valve opening is obtained. With gas flow The relationship.

[0025] S2. Establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill, and use an adaptive gradient-based regression identification algorithm to identify the gas flow rate-outlet temperature relationship model to obtain the parameters of the gas flow rate-outlet temperature relationship model. Optionally, the model establishing the relationship between the combustion gas flow rate and outlet temperature in the roller mill combustion chamber in step S2 specifically includes: The combustion chamber outlet temperature is related to the gas flow rate. At any given time, establish a display of the outlet temperature. With gas flow The gas flow rate-outlet temperature relationship model is as follows: (10) The gas flow rate-outlet temperature relationship model (10) is a typical first-order linear system with time-varying parameters. In model (10): The outlet temperature is expressed in units of... , These are slow, time-varying parameters, dimensionless. For bounded process noise, the unit is . .

[0026] Because the gas flow rate-outlet temperature relationship model (10) is a typical first-order linear system with time-varying parameters, the standard least squares algorithm cannot handle time-varying parameters and needs to introduce a forgetting factor to increase the weight of new data. However, the least squares algorithm is difficult to balance tracking speed and estimation accuracy and requires additional matrix operations. Therefore, in this embodiment of the invention, a regression identification algorithm based on adaptive gradient is proposed, which automatically adjusts the learning rate in different parameter directions while ensuring fast response.

[0027] Optionally, in step S2, an adaptive gradient-based regression identification algorithm is used to identify the gas flow rate-outlet temperature relationship model to obtain the parameters of the gas flow rate-outlet temperature relationship model, specifically including: Based on the parameter estimates from the previous moment Calculate residuals and parameters The instantaneous gradient; Calculation parameters The first moment is used to smooth the instantaneous gradient; Calculation parameters The second moment is used to constrain the scale of the instantaneous gradient; For parameters The first and second moments are corrected; The parameter estimates were obtained using the updated first and second moments. Update.

[0028] Optionally, in step S2, an adaptive gradient-based regression identification algorithm is used to identify the gas flow rate-outlet temperature relationship model to obtain the parameters of the gas flow rate-outlet temperature relationship model, specifically including: Take every moment Treating it as a mini-batch, applying gradient descent to the squared loss, and using adaptive second-order momentum to adjust the learning rate of each parameter to better track changes in time-varying parameters, the specific operations are as follows: Defined in At time , the parameter estimates are available. The residual at time is: (11) and parameters The instantaneous gradient is: (12) In the formula: The gradient is instantaneous and dimensionless. This is a quadratic loss, i.e., the minimum mean square error, which is dimensionless. Instantaneous gradients are easily affected by noise, and directly using instantaneous gradients will cause significant fluctuations in the identification results. Therefore, a weighted average of the instantaneous gradients is performed, as shown in Equation (13), which gives the first moment of the parameters. The gradient is smoothed to prevent it from oscillating drastically due to noise points. (13) In the formula: The first-order moment decay weight is dimensionless. In addition, the gradient may remain consistently large or small, which can compromise the convergence of the algorithm. Therefore, a weighted average of the squared gradients is needed to constrain the gradient scale, as shown in Equation (14), by setting the second moment. : (14) In the formula: The second-order moment decay weight is dimensionless. and Need to meet , (For estimating time-varying parameters, it is generally preferable to use smaller values.) ); In the early stages of iteration, due to the weighted average first moment and second moment The result will be close to 0. Therefore, as shown in equation (15), the first and second moments are modified to avoid unreasonable iterative updates in the initial stage of identification. (15) Finally, as shown in equation (16), the estimated parameters are updated using the updated first and second moments: (16) In the formula: The second-order moment decay weight is dimensionless (generally let...). , It is a constant. However, for estimating time-varying parameters, it is usually set to... The decay step size should be constant or bounded to ensure the algorithm's responsiveness and have... (steady-state error) It is a very small number that is close to 0 and prevents the numerator from being zero (generally taking a value of 0). ).

[0029] When the gradient is large for a long period of time, the iterative algorithm will automatically reduce the update magnitude to avoid parameter estimation oscillations; conversely, when the gradient is small for a long period of time, it will automatically increase the update magnitude to avoid iteration stagnation.

[0030] S3. Based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model, the outlet temperature is stably controlled by an autonomous valve opening adjustment algorithm.

[0031] Optionally, S3 specifically includes: S31. Based on the parameters of the valve opening-gas flow rate model and the parameters of the gas flow rate-outlet temperature model, the first valve opening is calculated using a feedforward control strategy. S32. Add temperature-level feedback control. By introducing proportional-integral control, the opening degree of the first valve is corrected in a closed loop to ensure the robustness and engineering feasibility of the outlet temperature control. S33. Add an accumulator to determine when to adjust the valve opening; S34. Adjust the valve opening to the corresponding physical range.

[0032] Optionally, S31 specifically includes: Assuming the target outlet temperature of the combustion chamber is Therefore, according to model (10), in order to achieve the target temperature, The required gas flow rate at any given time is: (17) Within the pressure range of the current pressure. Inside, the expected traffic Substituting into model (1), the required valve opening is obtained as follows: (18) Since the opening degree is fixed for each adjustment. Therefore, the valve opening control strategy is as follows: (19) (20) To ensure the appropriateness of the valve opening, it is necessary to determine whether the updated valve opening needs to be adjusted. Cut to the corresponding physical range (the valve opening after cutting is the same as the first valve opening), that is: (twenty one).

[0033] Equations (17)-(21) are the feedforward control strategies of this invention, when When the valve opening is within a small acceptable range, it indicates that the valve opening is within the desired range. However, this feedforward control method relies too heavily on model accuracy and ignores uncertainties caused by parameter estimation errors, measurement noise, and process noise. Especially in the complex industrial production environment of a pelletizing plant, if there is a significant deviation in parameter estimation or a decrease in the accuracy of a certain piece of equipment, the theoretically correct valve opening may actually lead to a significant temperature deviation. Therefore, the valve control algorithm in this embodiment of the invention also needs to incorporate temperature-level feedback control. By introducing proportional-integral control, the valve opening is further closed-loop corrected to ensure robustness and engineering feasibility in outlet temperature control.

[0034] Optionally, S32 specifically includes: definition The temperature error at any given time is: (twenty two) Then based on temperature error Design a PI control system: (twenty three) In the formula: This is the output for the proportional portion. For proportional gain, Output the integral part. For integral gain, The sampling period is dimensionless. Finally, the PI output is mapped proportionally to the theoretical opening increment: (twenty four) In the formula: It is an empirical coefficient that converts the outlet temperature difference into valve opening (its approximate value can be quickly obtained through a linear mapping of flow rate to valve opening), and it is dimensionless.

[0035] Optionally, S33 specifically includes: Because the valve opening change is fixed. Therefore, an accumulator is added to determine when the valve opening can be adjusted: Assuming in time, This indicates that the valve opening is within a small range of the theoretical valve opening. Further correction is then achieved through PI control, at which point an accumulator is introduced. Each time the PI controller output is received, the accumulator is updated: (25) Then, through judgment and The relationship between their sizes determines the valve opening adjustment command: like ,So , ;like ,So , ;like ,So and It remains unchanged.

[0036] Finally, adjust the valve opening to the corresponding physical range, that is... If the adjusted valve opening If the physical range is still exceeded, it indicates that the integral term has saturated, and the update of the integral term should be stopped. .

[0037] like Figure 2 As shown, this embodiment of the invention also provides a system for determining the relationship between the valve opening degree and the gas flow rate in the combustion chamber of a roller mill, characterized in that the system includes: The first module 210 is used to divide the fluctuation range of the combustion gas pressure in the roller mill combustion chamber into multiple intervals, and to establish a valve opening-gas flow relationship model in each interval. The model conversion module 220 is used to convert the valve opening degree-gas flow rate relationship model into a linear model in which the logarithm of the gas flow rate and the valve opening degree have a linear relationship. The second module 230 is used to establish a measurement value model to describe the saturation measurement characteristics of the sensor that observes the gas flow rate, and to establish a measurement flag vector for the gas flow rate measurement value. Frequency statistics module 240 is used to count the frequency of each dimension of the measurement mark vector being 1; The parameter identification module 250 is used to identify the parameters of the linear model based on the frequency using an empirical measure identification algorithm, and obtain estimated values ​​of the parameters of the linear model. The calculation and determination module 260 is used to calculate the parameters of the valve opening-gas flow relationship model based on the estimated values ​​of the parameters of the linear model, and to determine the relationship between the valve opening and gas flow in the combustion chamber of the roller mill based on the parameters of the valve opening-gas flow relationship model.

[0038] like Figure 3 As shown, this embodiment of the invention also provides a combustion chamber outlet temperature control system for a roller mill, the system comprising: The first identification module 310 is used to establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model. The second identification module 320 is used to establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill. It uses an adaptive gradient-based regression identification algorithm to identify the gas flow rate-outlet temperature relationship model and obtain the parameters of the gas flow rate-outlet temperature relationship model. The stability control module 330 is used to stabilize the outlet temperature by autonomously adjusting the valve opening algorithm based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model.

[0039] The combustion chamber outlet temperature control system for a roller mill provided in this embodiment of the invention has a functional structure that corresponds to the combustion chamber outlet temperature control method for a roller mill provided in this embodiment of the invention, and will not be described again here.

[0040] Figure 4 This is a schematic diagram of the structure of an electronic device 400 provided in an embodiment of the present invention. The electronic device 400 may vary considerably due to different configurations or performance. It may include one or more central processing units (CPUs) 401 and one or more memories 402. The memory 402 stores at least one instruction, which is loaded and executed by the processor 401 to implement the steps of the above-described method for controlling the outlet temperature of the combustion chamber of a roller mill.

[0041] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including instructions that can be executed by a processor in a terminal to perform the above-described method for controlling the outlet temperature of the roller mill combustion chamber. For example, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device.

[0042] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0043] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for controlling the outlet temperature of a roller mill combustion chamber, characterized in that, The method includes: S1. Establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model; S2. Establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill, and use an adaptive gradient-based regression identification algorithm to identify the gas flow rate-outlet temperature relationship model to obtain the parameters of the gas flow rate-outlet temperature relationship model. S3. Based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model, the outlet temperature is stably controlled by an autonomous valve opening adjustment algorithm.

2. The method according to claim 1, characterized in that, S1 specifically includes: exist At that moment, the valve opening degree was So when the gas pressure In the interval , At that time, the following valve opening-gas flow relationship model is established: （1） In model (1): For valve adjustable ratio, The maximum valve opening is dimensionless. The valve opening degree-gas flow rate relationship model is identified to obtain the parameters of the valve opening degree-gas flow rate relationship model.

3. The method according to claim 2, characterized in that, The S2 section establishes a model for the relationship between the combustion gas flow rate and the outlet temperature in the roller mill combustion chamber, specifically including: The combustion chamber outlet temperature is related to the gas flow rate. At any given time, establish a display of the outlet temperature. With gas flow The gas flow rate-outlet temperature relationship model is as follows: （10） The gas flow rate-outlet temperature relationship model (10) is a typical first-order linear system with time-varying parameters. In model (10): The outlet temperature is expressed in units of... , These are slow, time-varying parameters, dimensionless. For bounded process noise, the unit is . .

4. The method according to claim 3, characterized in that, In step S2, an adaptive gradient-based regression identification algorithm is used to identify the gas flow rate-outlet temperature relationship model, obtaining the parameters of the gas flow rate-outlet temperature relationship model, specifically including: Based on the parameter estimates from the previous moment Calculate residuals and parameters The instantaneous gradient; Calculation parameters The first moment is used to smooth the instantaneous gradient; Calculation parameters The second moment is used to constrain the scale of the instantaneous gradient; For parameters The first and second moments are corrected; The parameter estimates were obtained using the updated first and second moments. Update.

5. The method according to claim 4, characterized in that, In step S2, an adaptive gradient-based regression identification algorithm is used to identify the gas flow rate-outlet temperature relationship model, obtaining the parameters of the gas flow rate-outlet temperature relationship model, specifically including: Take every moment Treating it as a mini-batch, applying gradient descent to the squared loss, and using adaptive second-order momentum to adjust the learning rate of each parameter to better track changes in time-varying parameters, the specific operations are as follows: Defined in At time , the parameter estimates are available. The residual at time is: （11） and parameters The instantaneous gradient is: （12） In the formula: The gradient is instantaneous and dimensionless. This is a quadratic loss, i.e., the minimum mean square error, which is dimensionless. Instantaneous gradients are easily affected by noise, and directly using instantaneous gradients will cause significant fluctuations in the identification results. Therefore, a weighted average of the instantaneous gradients is performed, as shown in Equation (13), which gives the first moment of the parameters. The gradient is smoothed to prevent it from oscillating drastically due to noise points. （13） In the formula: The first-order moment decay weight is dimensionless. In addition, the gradient may remain consistently large or small, which can compromise the convergence of the algorithm. Therefore, a weighted average of the squared gradients is needed to constrain the gradient scale, as shown in Equation (14), by setting the second moment. : （14） In the formula: The second-order moment decay weight is dimensionless. and Need to meet , ; In the early stages of iteration, due to the weighted average first moment and second moment The result will be close to 0. Therefore, as shown in equation (15), the first and second moments are modified to avoid unreasonable iterative updates in the initial stage of identification. （15） Finally, as shown in equation (16), the estimated parameters are updated using the updated first and second moments: （16） In the formula: The second-order moment decay weight is dimensionless. It is to prevent the numerator from being zero, which is a very small number close to 0.

6. The method according to claim 5, characterized in that, S3 specifically includes: S31. Based on the parameters of the valve opening-gas flow rate model and the parameters of the gas flow rate-outlet temperature model, the first valve opening is calculated using a feedforward control strategy. S32. Add temperature-level feedback control. By introducing proportional-integral control, the opening degree of the first valve is corrected in a closed loop to ensure the robustness and engineering feasibility of the outlet temperature control. S33. Add an accumulator to determine when to adjust the valve opening; S34. Adjust the valve opening to the corresponding physical range.

7. The method according to claim 6, characterized in that, S31 specifically includes: Assuming the target outlet temperature of the combustion chamber is Therefore, according to model (10), in order to achieve the target temperature, The required gas flow rate at any given time is: （17） Within the pressure range of the current pressure. Inside, the expected traffic Substituting into model (1), the required valve opening is obtained as follows: （18） Since the opening degree is fixed for each adjustment. Therefore, the valve opening control strategy is as follows: （19） （20） To ensure the appropriateness of the valve opening, it is necessary to determine whether the updated valve opening needs to be adjusted. Cut to the corresponding physical area, that is: （21）。 8. The method according to claim 7, characterized in that, S32 specifically includes: definition The temperature error at any given time is: （22） Then based on temperature error Design a PI control: （23） In the formula: This is the output for the proportional portion. For proportional gain, Output the integral part. For integral gain, The sampling period is dimensionless. Finally, the PI output is mapped proportionally to the theoretical opening increment: （24） In the formula: It is an empirical coefficient that converts the outlet temperature difference into valve opening degree, and it is dimensionless.

9. The method according to claim 7, characterized in that, S33 specifically includes: Because the valve opening change is fixed. Therefore, an accumulator is added to determine when the valve opening can be adjusted: Assuming in time, This indicates that the valve opening is within a small range of the theoretical valve opening. Further correction is then achieved through PI control, at which point an accumulator is introduced. Each time the PI controller output is received, the accumulator is updated: （25） Then, through judgment and The relationship between their sizes determines the valve opening adjustment command: like ,So , ;like ,So , ;like ,So and It remains unchanged.

10. A combustion chamber outlet temperature control system for a roller mill, characterized in that, The system includes: The first identification module is used to establish a valve opening degree-gas flow rate relationship model for the combustion chamber of the roller mill, identify the valve opening degree-gas flow rate relationship model, and obtain the parameters of the valve opening degree-gas flow rate relationship model. The second identification module is used to establish a gas flow rate-outlet temperature relationship model for the combustion chamber of the roller mill. The model is identified using an adaptive gradient-based regression identification algorithm to obtain the parameters of the gas flow rate-outlet temperature relationship model. The stability control module is used to stabilize the outlet temperature by autonomously adjusting the valve opening algorithm based on the parameters of the valve opening-gas flow model and the parameters of the gas flow-outlet temperature model.

Images