A coal injection optimization control method based on Gaussian process regression

Abstract

The present invention discloses a coal injection optimization control method based on Gaussian process regression, including collecting adjustment parameters in the coal injection process for preprocessing; selecting parameters with high correlation with coal injection amount from the preprocessed parameters as characteristic values; using Gaussian process regression to train parameters and obtain a prediction model; saving the prediction model as a reloadable file, and inputting real-time data for prediction. The present invention reduces the phenomenon of abnormal fluctuations in coal injection amount caused by human intervention for the manual adjustment of the coal injection system mode, improves the stability of uniform coal injection, optimizes the blast furnace condition, and improves ironmaking efficiency and quality; for new employees, in the absence of relevant ironmaking experience, the purpose of optimizing the blast furnace coal injection operation mode is to provide decision support to on-site operators, play a role in assisting learning, and promote experience growth. At the same time, it also plays a role in promoting the standardized operation of coal injection workers and improving the operation level.

Description

Technical Field

[0001] The invention relates to the technical field of coal injection in blast furnace ironmaking, and in particular to a coal injection optimization control method based on Gaussian process regression. Background Art

[0002] Coal injection in blast furnaces is an important technical means in current blast furnace ironmaking. It refers to the direct injection of finely ground coal into the furnace from the blast furnace tuyere to replace the more expensive coke to provide heat and reducing agent to the blast furnace. Coal injection in blast furnaces can reduce the coke ratio and significantly reduce the cost of blast furnace ironmaking. At the same time, it can also adjust the furnace condition, improve the working state of the furnace, and stabilize the operation of the blast furnace. At present, there are two main operating modes of the coal injection system of the blast furnace:

[0003] The first mode is: manual mode to adjust the coal injection system; the manual adjustment of the coal injection system is that workers adjust the coal injection control quantity (such as tank pressure, mixing pressure, etc.) according to experience to achieve control of the coal injection flow rate. It is impossible to perform optimal precise adjustment, and it requires high experience from the job operators. The adjustment error in this mode is large, and large fluctuations in the coal injection flow rate are likely to occur, thereby causing furnace temperature fluctuations, affecting the stability of the blast furnace condition, and is not conducive to further increasing the blast furnace output. This mode is widely used in coal injection systems that discharge coal from the upper part of the injection tank.

[0004] The second mode is: automatic adjustment of coal injection system; the blast furnace adopts automatic coal injection technology to achieve continuous and uniform injection volume of the blast furnace, so that the furnace condition reaches a relatively stable equilibrium state, which can increase the output of the blast furnace and increase profits for the enterprise. At present, this mode is mainly used in the coal injection system with coal discharged from the lower part of the injection tank. In recent years, the newly built large steel mills have basically adopted this model to achieve automatic coal injection, but most of the existing steel mills are limited by the equipment conditions of the upper coal discharge of the injection tank, and mainly use manual adjustment mode. The process equipment used in these two coal injection systems is quite different. It is difficult to achieve automation through the transformation of process equipment, and the transformation cost required is also relatively large. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed.

[0007] Therefore, the present invention provides a coal injection optimization control method based on Gaussian process regression, which can realize blast furnace ironmaking coal injection optimization control for a coal injection system under manual mode conditions.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: including, collecting the adjustment parameters in the coal injection process for preprocessing; selecting parameters with a high correlation with the coal injection amount from the preprocessed parameters as characteristic values; using Gaussian process regression to perform parameter training to obtain a prediction model; saving the prediction model as a file that can be reloaded, and inputting real-time data for prediction.

[0009] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, the collection includes collecting historical operation data with high relevance from the blast furnace ironmaking database, obtaining the characteristic value data after preprocessing, and taking the actual coal injection flow rate as the target value.

[0010] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, the characteristic values ​​include tank weight, tank pressure, mixing pressure, air supplement flow, middle fluidization flow, and lower fluidization flow.

[0011] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, wherein: the functional relationship about the coal injection flow rate is formed by using the characteristic value, including:

[0012] f(flow)=f(W,P,P mix ,F air ,F mid ,F btm )

[0013] Where W is the tank weight, P is the tank pressure, P mix Mixed pressure, F air is the air supply flow rate, F mid is the middle fluidization flow rate, F btm is the lower fluidization flow rate, and flow is the actual coal injection flow rate.

[0014] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, wherein: the training utilizes Gaussian process regression to fit the characteristic value and the target value in the functional relationship.

[0015] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, the effect of the prediction model training is constrained by influencing factors, which include the size of the training set, the training data itself and the selected Gaussian kernel function.

[0016] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, the Gaussian kernel function includes radial basis function, exponential function kernel and rational quadratic function kernel.

[0017] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, it also includes: comparing the root mean square error obtained by training the prediction model, and taking it as an indicator of the training effect. The smaller the root mean square error value is, the better the fitting effect is. After the training is completed, the corresponding prediction model is derived.

[0018] As a preferred solution of the coal injection optimization control method based on Gaussian process regression described in the present invention, it includes: loading the trained prediction model, connecting to the blast furnace ironmaking database, obtaining real-time eigenvalue data as input of the prediction model, and the goal is to obtain the optimized objective function,

[0019] ∑=α1·|Δflow|+α2·|ΔP|+α3·|ΔP mix |

[0020] Among them, α i is the weighting coefficient.

[0021] As a preferred scheme of the coal injection optimization control method based on Gaussian process regression described in the present invention, it also includes: exhaustively enumerating and weighted processing the characteristic value data, adjusting the current input parameters, and calculating the optimal output corresponding to the current input; selecting two pressure parameters that have a relatively large impact on the coal injection flow rate and enumerating them within the range of ±5kPa; inputting the obtained data group into the prediction model to obtain the corresponding predicted value group of the coal injection flow rate; according to the above-mentioned objective function, subtracting the coal injection flow rate in each group of data after exhaustion from the set flow rate, and subtracting the tank pressure and the mixed pressure from the corresponding real-time data; performing weighted summation on the three listed differences to obtain the minimum value after the weighted summation, and the corresponding control parameter is the optimal control parameter, and its control quantity changes little, the system fluctuation is small, and the flow rate approaches the set value.

[0022] The beneficial effects of the present invention are as follows: the present invention aims at manually adjusting the coal injection system mode, reduces the phenomenon of abnormal fluctuation of coal injection amount caused by human intervention, improves the stability of uniform coal injection, optimizes blast furnace conditions, and improves ironmaking efficiency and quality; for new employees, in the absence of relevant ironmaking experience, the purpose of optimizing the blast furnace coal injection operation mode is to provide decision-making support to on-site operators, play a role in assisting learning, and promoting experience growth, and at the same time, it also plays a role in promoting the standardized operation of coal injection workers and improving the operation level. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0024] Figure 1 It is a flow chart of the coal injection optimization control method based on Gaussian process regression according to the first embodiment of the present invention;

[0025] Figure 2 A schematic diagram of a blast furnace ironmaking coal injection process according to a coal injection optimization control method based on Gaussian process regression according to a first embodiment of the present invention;

[0026] Figure 3 A schematic diagram of the training and calling process of the coal injection flow prediction model of the coal injection optimization control method based on Gaussian process regression according to the first embodiment of the present invention;

[0027] Figure 4 It is a schematic diagram for comparing stability curves of the coal injection optimization control method based on Gaussian process regression described in the second embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0030] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0031] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0032] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0033] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0034] Example 1

[0035] Reference Figure 1 to Figure 3 , which is the first embodiment of the present invention, provides a coal injection optimization control method based on Gaussian process regression, comprising:

[0036] S1: Collect the adjustment parameters in the coal injection process for preprocessing. It should be noted that:

[0037] The collection includes collecting historical operation data with high relevance from the blast furnace ironmaking database, obtaining characteristic value data after preprocessing, and taking the actual coal injection flow rate as the target value.

[0038] S2: Select the parameters with high correlation with coal injection amount from the pre-processed parameters as characteristic values. What needs to be explained in this step is:

[0039] The characteristic values ​​include tank weight, tank pressure, mixing pressure, air supply flow, middle fluidization flow, and lower fluidization flow;

[0040] The functional relationship about coal injection flow rate is formed by using characteristic values, including:

[0041] f(flow)=f(W,P,P mix ,F air ,F mid ,F btm )

[0042] Where W is the tank weight, P is the tank pressure, P mix Mixed pressure, F air is the air supply flow rate, F mid is the middle fluidization flow rate, F btm is the lower fluidization flow rate, and flow is the actual coal injection flow rate.

[0043] S3: Use Gaussian process regression to train parameters and obtain a prediction model. Figure 3 , which also needs to be explained is:

[0044] The training uses Gaussian process regression to fit the eigenvalues ​​and target values ​​in the functional relationship;

[0045] The effectiveness of predictive model training is constrained by influencing factors, including the size of the training set, the training data itself, and the selected Gaussian kernel function;

[0046] Gaussian kernel functions include radial basis function, exponential function kernel and rational quadratic function kernel;

[0047] Compare the root mean square error obtained by training the prediction model and use it as an indicator of the training effect. The smaller the root mean square error value, the better the fitting effect. After the training is completed, the corresponding prediction model is derived;

[0048] Load the trained prediction model, connect to the blast furnace ironmaking database, and obtain real-time eigenvalue data as the input of the prediction model. The goal is to obtain the optimized objective function.

[0049] ∑=α1·|Δflow|+α2·|ΔP|+α3·|ΔP mix |

[0050] Among them, α i is the weighting coefficient.

[0051] S4: Save the prediction model as a file that can be reloaded, and input real-time data for prediction.

[0052] Perform exhaustive and weighted processing on the eigenvalue data, adjust the current input parameters, and calculate the optimal output corresponding to the current input;

[0053] Two pressure parameters with relatively large influence on coal injection flow rate were selected for exhaustive enumeration within the range of ±5kPa;

[0054] Input the obtained data group into the prediction model to obtain the corresponding predicted value group of coal injection flow rate;

[0055] According to the above objective function, the coal injection flow rate in each set of data after exhaustive enumeration is subtracted from the set flow rate, and the tank pressure and mixed pressure are subtracted from the corresponding real-time data;

[0056] The three listed differences are weighted summed to obtain the smallest value after the weighted summation. The corresponding control parameter is the optimal control parameter, with small changes in the control quantity, small system fluctuations, and the flow rate approaching the set value.

[0057] Generally speaking, in the blast furnace coal injection system, the coal injection flow control is determined by multiple variables. There are multiple variables and they influence each other. The strength of each condition variable is also different. Figure 2 It mainly stabilizes the tank pressure and controls the coal injection flow rate by adjusting the valve associated with the injection tank. It comprehensively considers the adjustment parameters in the coal injection process and selects the parameters with high correlation with the coal injection amount as the characteristic values ​​for modeling.

[0058] Preferably, a large amount of process data will be generated during the actual operation of the blast furnace coal injection system. Due to external interference and system factors, some incomplete or erroneous data will inevitably exist. If these data are used, the accuracy and precision of the production process prediction model will be seriously affected, which is not conducive to the continuous optimization of the model. Therefore, it is necessary to select and preprocess the production data with eigenvalues ​​to provide relatively complete data for modeling and optimization of the production process.

[0059] It is not difficult to understand that by selecting relatively complete feature values, a prediction model is obtained through data training, and the model is saved as a file that can be reloaded. This embodiment mainly trains the model on historical data and predicts the real-time data based on the saved model to achieve optimized control of coal injection in blast furnace ironmaking.

[0060] Reference Figure 2 , is a schematic diagram of the coal injection process for blast furnace ironmaking. In order to keep the blast furnace condition as stable as possible, the coal powder injection rate is controlled to be kept in a uniform state by adjusting the pressure of the injection tank and the injection pipeline in the coal injection system. Among them, adjusting the #2 charging valve and the #3 pressure-making valve can control the tank pressure of the injection tank, adjusting the #4 middle fluidizing valve and the #5 bottom fluidizing valve can control the fluidity of the coal powder in the injection tank, adjusting the #6 coal outlet valve and the #7 coal supply valve can control the coal powder injection amount, and adjusting the #8 air supply valve can control the mixing pressure of the injection pipeline and the coal powder injection rate.

[0061] It should also be noted that in this embodiment, based on the Gaussian process regression model, by selecting different Gaussian kernel functions, the historical characteristic value data related to ironmaking production are trained and modeled to generate an callable optimization model, and the model is periodically updated. Then, by calling the established prediction model, the real-time ironmaking data is processed to realize the prediction function of coal injection flow rate and provide decision support for correlation parameter adjustment. Through the intelligent optimization control method of blast furnace ironmaking coal injection based on the Gaussian process regression model, the effective utilization rate of coal powder is improved, the blast furnace condition is stabilized, and the purpose of optimizing the production indicators of the blast furnace system is achieved.

[0062] This method takes the coal injection amount of blast furnace as the research object, based on a large amount of actual industrial operation data of blast furnace in steel plant, and uses Gaussian process regression model combined with important feature data sets selected by blast furnace experts for model training. This method does not need to add external hardware equipment, and effectively utilizes a large amount of process data generated during the blast furnace smelting process. After model training, the function of optimizing coal injection control in blast furnace ironmaking is finally realized; on most of the currently built coal injection systems with coal discharged from the upper part of the injection tank, without adding external hardware equipment, the injection tank can discharge coal and inject evenly and stably according to the set coal injection amount, so as to minimize the fluctuation of furnace conditions as much as possible.

[0063] Example 2

[0064] Reference Figure 4 , which is the second embodiment of the present invention. This embodiment is different from the first embodiment in that it provides a verification of a coal injection optimization control method based on Gaussian process regression, including:

[0065] In order to better verify and illustrate the technical effects of the method of the present invention, this embodiment selects the traditional manual adjustment method of the coal injection system and the method of the present invention for comparative testing, compares the test results by scientific demonstration means, and verifies the real effect of the method of the present invention.

[0066] The traditional manual adjustment method of the coal injection system cannot perform the most optimized precise adjustment, and has high requirements on the experience of the operators. The adjustment error in this mode is large, and the coal injection flow rate is prone to large fluctuations, which causes furnace temperature fluctuations and affects the stability of the blast furnace condition, which is not conducive to further improving the blast furnace output. In order to verify that the method of the present invention has higher prediction accuracy and stability in predicting and adjusting the coal injection system than the traditional method, this embodiment will use the traditional method and the method of the present invention to respectively perform real-time measurement and comparison of the coal injection flow rate of the simulated coal injection system.

[0067] Test environment: The simulated coal injection system is run on a simulation platform to simulate the operation and simulate the coal injection scenario. Historical coal injection correlation data are used as test samples. The traditional method of manual adjustment operation is used to perform prediction tests and obtain test results. The method of the present invention is used to import the prediction model program and use MATLB to implement the simulation test of the method of the present invention, and the simulation data is obtained according to the experimental results. Each method tests 100 groups of data, calculates the time and the root mean square of the prediction error for each group of data, and compares the error calculation with the actual prediction value of the simulation input.

[0068] Reference Figure 4 The solid line is the curve output by the method of the present invention, and the dotted line is the curve output by the traditional method. Figure 4 It can be seen intuitively that the solid line and the dotted line show different trends with the increase of time. Compared with the dotted line, the solid line has been showing a stable upward trend in the early stage. Although it has declined in the later stage, the fluctuation is not large and it has been above the dotted line and maintained a certain distance. The dotted line shows a large fluctuation trend and is unstable. Therefore, the efficiency of the solid line is always greater than that of the dotted line, which verifies the real effect of the method of the present invention.

[0069] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A coal injection optimization control method based on Gaussian process regression, characterized in that: include, Collect the adjustment parameters in the coal injection process for preprocessing; Selecting parameters with a high correlation with the coal injection amount from the pre-processed parameters as characteristic values; Gaussian process regression is used to train parameters and obtain a prediction model; The prediction model is saved as a file that can be reloaded, and real-time data is input for prediction; The collection includes collecting historical operation data with high relevance from a blast furnace ironmaking database, obtaining the characteristic value data after preprocessing, and taking the actual coal injection flow rate as the target value; The characteristic values ​​include tank weight, tank pressure, mixing pressure, air supply flow, middle fluidization flow, and lower fluidization flow; The functional relationship about the coal injection flow rate is formed by using the characteristic value, including: f(flow)=f(W,P,P mix ,F air ,F mid ,F btm ) Where W is the tank weight, P is the tank pressure, P mix is the mixed pressure, F air is the air supply flow rate, F mid is the middle fluidization flow rate, F btm is the lower fluidization flow rate, flow is the actual coal injection flow rate; Compare the root mean square error obtained by training the prediction model and use it as an indicator of the training effect. The smaller the root mean square error value is, the better the fitting effect is. After the training is completed, the corresponding prediction model is derived; Load the trained prediction model, connect to the blast furnace ironmaking database, and obtain real-time feature value data as the input of the prediction model. The goal is to obtain the optimized objective function. ∑=α1·|Δflow|+α2·|ΔP|+α3·|ΔP mix | Among them, α i is the weighting coefficient; Performing exhaustive and weighted processing on the eigenvalue data, adjusting the current input parameters, and calculating the optimal output corresponding to the current input; Select two pressure parameters that have a relatively large impact on the coal injection flow rate and perform exhaustive enumeration within the range of ±5 kPa; Inputting the obtained data group into the prediction model to obtain the corresponding predicted value group of the coal injection flow rate; According to the above objective function, the coal injection flow rate in each set of data after exhaustive enumeration is subtracted from the set flow rate, and the tank pressure and mixed pressure are subtracted from the corresponding real-time data; The three listed differences are weighted summed to obtain the smallest value after the weighted summation. The corresponding control parameter is the optimal control parameter, with small changes in the control quantity, small system fluctuations, and the flow rate approaching the set value.

2. The coal injection optimization control method based on Gaussian process regression according to claim 1 is characterized in that: The training utilizes Gaussian process regression to fit the characteristic value and the target value in the functional relationship.

3. The coal injection optimization control method based on Gaussian process regression according to claim 1 or 2, characterized in that: The effect of the prediction model training is constrained by influencing factors, which include the size of the training set, the training data itself, and the selected Gaussian kernel function.

4. The coal injection optimization control method based on Gaussian process regression according to claim 3 is characterized in that: The Gaussian kernel functions include radial basis functions, exponential function kernels and rational quadratic function kernels.

Images