Converter flue gas component analysis method based on AI model

Abstract

The invention provides a converter flue gas component analysis method based on an AI model, and the method comprises the following steps: S1, collecting historical detection data of converter flue gas, and processing the data; s2, constructing a converter flue gas analysis model by using the processed converter flue gas historical data; s3, collecting a current spectral image of the converter through non-dispersive infrared detection equipment; s4, preprocessing the current spectral image of the converter, and inputting the preprocessed current spectral image into the converter flue gas analysis model to obtain a current converter flue gas component result; converter flue gas components can be quickly analyzed according to detection data collected by non-dispersive infrared detection equipment by constructing a converter flue gas analysis model, so that the converter process is accurately controlled, inclusions and defects in steel are reduced, higher-quality steel is produced, and the requirement of the market for high-quality steel is met.

Description

Technical Field

[0001] The present invention relates to the technical field of converter steelmaking detection, and particularly to a method for analyzing the components of converter flue gas based on an AI model. Background Art

[0002] Converter steelmaking uses molten iron, scrap steel, and ferroalloys as the main raw materials. Without relying on external energy, the steelmaking process is completed in a converter by the physical heat of the molten iron itself and the chemical reactions between the components of the molten iron. Converters are classified into acidic and basic according to the refractory materials, and into top-blowing, bottom-blowing, and side-blowing according to the position where the gas is blown into the furnace; they are classified into air converters and oxygen converters according to the type of gas. Basic oxygen top-blowing and top-bottom combined blowing converters are the most commonly used steelmaking equipment due to their high production speed, large output, high single-furnace output, low cost, and low investment.

[0003] With the development of science and technology, converter steelmaking technology is also shifting towards automated and intelligent production. Among them, converter flue gas detection is an important part of intelligent production. By grasping the flue gas components in real time, especially the contents of oxygen, carbon monoxide, carbon dioxide, etc., it provides a basis for accurately controlling the oxygen blowing time, intensity, and rhythm, which can improve the steelmaking efficiency and the quality of molten steel; at the same time, by detecting the flue gas components, the progress and status of the chemical reactions in the furnace can be grasped, abnormalities can be discovered in time, and the feeding and operation parameters can be adjusted to ensure the stable operation of the converter; in addition, converter flue gas detection can understand the content of harmful impurities in the flue gas, and then take measures to reduce the generation of impurities and improve the purity of molten steel; by analyzing the flue gas components, the energy supply and combustion conditions can be reasonably adjusted to reduce energy consumption and greenhouse gas emissions, achieving the effect of energy conservation and emission reduction; however, there is currently a lack of a solution for accurately analyzing the components of converter flue gas in the existing technology.

[0004] Therefore, there is an urgent need for a method for analyzing the components of converter flue gas based on an AI model, which can solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for analyzing the components of converter flue gas based on an AI model to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above purpose, the present invention provides the following solutions:

[0007] The present invention provides a method for analyzing the components of converter flue gas based on an AI model, including the following steps:

[0008] S1. Collect historical detection data of converter flue gas and process the data;

[0009] S2. Use the processed historical data of converter flue gas to construct a converter flue gas analysis model;

[0010] S3. Collect the current spectral image of the converter through a non-dispersive infrared detection device;

[0011] S4. Input the preprocessed current spectral image of the converter into the converter flue gas analysis model to obtain the current converter flue gas composition result.

[0012] Preferably, in step S1, the historical converter flue gas detection data includes the type of flue gas, the concentration of flue gas, and the non-dispersive infrared detection data of the flue gas.

[0013] Preferably, in step S1, the method for processing the data includes:

[0014] S11. Denoise the spectral image of the non-dispersive infrared detection data of the flue gas;

[0015] S12. Calculate the infrared absorption spectrum curve of the flue gas using the flue gas spectral inversion algorithm.

[0016] Preferably, in step S11, the denoising method uses the following formula:

[0017]

[0018] where i represents the coordinates (X, Y, Z) in the data cube, where X and Y represent the abscissa and ordinate in the spatial domain respectively, and Z represents the spectral dimension coordinate; N(i) represents the neighborhood of i, represents the gray value of i; the weight coefficient reflects the similarity between y i and y j .

[0019] Preferably, in step S12, the flue gas spectral inversion algorithm uses the following formula:

[0020] S calc =(1 - τ gas )(B gas τ air + S2 - S1);

[0021]

[0022] where τ gas represents the transmittance of the measured gas, B gas represents the Planck radiation of the measured gas, τ air represents the transmittance of the air between the measured gas and the non-dispersive infrared detection device, S calc represents the calculated value of the radiation of the measured gas, S2 represents the radiation of the air between the measured gas and the non-dispersive infrared detection device, S1 represents the radiation of the air between the background and the non-dispersive infrared detection device, α iThe absorption rate of the \(i\)-th gas in the gas spectral database for light with a frequency of \(v\) is denoted as \((ν)\), \(C\) i is the concentration of the gas, and \(l\) is the optical path of light propagating in the gas.

[0023] Preferably, in step S2, the method for constructing the converter flue gas analysis model includes:

[0024] S21. Fusing the concentration data of carbon dioxide gas and carbon monoxide gas with the flue gas infrared absorption spectrum curve to obtain an AI sub-model;

[0025] S22. Constructing an oxygen gas concentration calculation model;

[0026] S23. Using the detected concentration of hydrogen gas as an independent coefficient to fuse with the AI sub-model and the oxygen gas concentration calculation model to obtain the converter flue gas analysis model.

[0027] Preferably, in step S3, the non-dispersive infrared detection device includes an infrared light source. On the opposite side of the infrared light source, there are a non-dispersive infrared sensor and a semiconductor sensor. On the side close to the infrared light source between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor, there is an optical filter. On the side close to the sensor between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor, there is a narrowband filter.

[0028] Preferably, in step S4, the preprocessing includes signal conversion, signal amplification, and signal filtering.

[0029] The present invention has achieved the following beneficial technical effects compared with the prior art:

[0030] A converter flue gas composition analysis method based on an AI model provided by the present invention includes the following steps: S1. Collecting historical detection data of converter flue gas and processing the data; S2. Using the processed historical data of converter flue gas to construct a converter flue gas analysis model; S3. Collecting the current spectral image of the converter through a non-dispersive infrared detection device; S4. Inputting the preprocessed current spectral image of the converter into the converter flue gas analysis model to obtain the current converter flue gas composition result; By constructing the converter flue gas analysis model, the composition of the converter flue gas can be quickly analyzed based on the detection data collected by the non-dispersive infrared detection device, thereby precisely controlling the converter process, helping to reduce inclusions and defects in steel, and producing higher-quality steel to meet the market demand for high-quality steel. Description of the Drawings

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of a converter flue gas composition analysis method based on an AI model provided by the present invention. Detailed implementation manners

[0033] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0034] The purpose of the present invention is to provide a converter flue gas composition analysis method based on an AI model to solve the problems existing in the prior art.

[0035] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0036] Embodiment 1:

[0037] This embodiment provides a converter flue gas composition analysis method based on an AI model. As Figure 1 shown, it includes the following steps:

[0038] S1. Collect historical detection data of converter flue gas. The historical detection data of converter flue gas includes flue gas type, flue gas concentration, and non-dispersive infrared detection data of flue gas, and process the data. The methods for processing the data include:

[0039] S11. Perform noise reduction processing on the spectral image of the non-dispersive infrared detection data of the flue gas;

[0040] S12. Calculate the infrared absorption spectrum curve of the flue gas using the flue gas spectrum inversion algorithm.

[0041] The method for noise reduction adopts the following formula:

[0042]

[0043] where i represents the coordinates (X, Y, Z) in the data cube, where X and Y respectively represent the abscissa and ordinate in the spatial domain, and Z represents the spectral dimension coordinate; N(i) represents the neighborhood of i, represents the grayscale value of i; the weight coefficient reflects y i and y j the degree of similarity between;

[0044] The flue gas spectral inversion algorithm uses the following formula:

[0045] S calc =(1 - τ gas )(B gas τ air + S2 - S1);

[0046]

[0047] where τ gas represents the transmittance of the measured gas, B gas represents the Planck radiation of the measured gas, τ air represents the transmittance of the air between the measured gas and the non-dispersive infrared detection device, S calc represents the calculated value of the radiation of the measured gas, S2 represents the radiation of the air between the measured gas and the non-dispersive infrared detection device, S1 represents the radiation of the air between the background and the non-dispersive infrared detection device, α i (ν) represents the absorption rate of the i-th gas in the gas spectral database for light with a frequency of v, C i is the concentration of the gas, and l is the optical path of the light propagating in the gas

[0048] S2. Construct a converter flue gas analysis model using the processed historical data of the converter flue gas; The method for constructing the converter flue gas analysis model includes:

[0049] S21. Fuse the concentration data of carbon dioxide gas and carbon monoxide gas with the flue gas infrared absorption spectral curve to obtain an AI sub-model;

[0050] S22. Construct an oxygen gas concentration calculation model;

[0051] S23. Fuse the detected concentration of hydrogen gas as an independent coefficient with the AI sub-model and the oxygen gas concentration calculation model to obtain the converter flue gas analysis model;

[0052] S3. Collect the current spectral image of the converter through a non-dispersive infrared detection device; The non-dispersive infrared detection device includes an infrared light source, a non-dispersive infrared sensor and a semiconductor sensor are provided on the opposite side of the infrared light source, an optical filter is provided on the side close to the infrared light source between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor, and a narrowband filter is provided on the side close to the sensor between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor;

[0053] S4. Input the current spectral image of the converter after preprocessing into the converter flue gas analysis model to obtain the current converter flue gas composition results. Among them, the preprocessing includes signal conversion, signal amplification, and signal filtering.

[0054] A converter flue gas composition analysis method based on an AI model provided by the present invention can quickly analyze the converter flue gas composition according to the detection data collected by the non-dispersive infrared detection device by constructing a converter flue gas analysis model, thereby precisely controlling the converter process, helping to reduce inclusions and defects in steel, producing higher-quality steel, and meeting the market demand for high-quality steel.

[0055] The present invention uses specific examples to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A converter flue gas composition analysis method based on an AI model, characterized in that: The following steps are involved: S1. Collect historical converter flue gas detection data and process the data; S2. construct a converter flue gas analysis model using the processed converter flue gas historical data; S3. Collecting the current spectral image of the converter through a non-dispersive infrared detection device; S4. The current spectral image of the converter is preprocessed and then input into the converter flue gas analysis model to obtain the current converter flue gas composition result.

2. The converter flue gas composition analysis method based on the AI ​​model according to claim 1 is characterized in that: In step S1, the converter flue gas historical detection data includes flue gas type, flue gas concentration and flue gas non-dispersive infrared detection data.

3. The converter flue gas composition analysis method based on the AI ​​model according to claim 2 is characterized in that: In step S1, the method for processing data includes: S11. performing noise reduction processing on the spectral image of the smoke non-dispersive infrared detection data; S12. Calculate the infrared absorption spectrum curve of the flue gas using the flue gas spectrum inversion algorithm.

4. The converter flue gas composition analysis method based on the AI ​​model according to claim 3 is characterized in that: In step S11, the noise reduction method adopts the following formula: Where i represents the coordinates (X, Y, Z) in the data cube, where X and Y represent the horizontal and vertical coordinates of the spatial domain, respectively, and Z represents the spectral dimension coordinate; N(i) represents the neighborhood of i, Represents the gray value of i; weight coefficient Reflects y i and j The degree of similarity between them.

5. The method for analyzing converter flue gas composition based on an AI model according to claim 3 is characterized in that: In step S12, the smoke spectrum inversion algorithm adopts the following formula: S calc =(1-τ gas )(B gas t air +S2-S1); Among them, τ gas Indicates the permeability of the gas being measured, B gas represents the Planck radiation of the measured gas, τ air Indicates the air transmittance between the measured gas and the non-dispersive infrared detection equipment, S calc represents the calculated value of the measured gas radiation, S2 represents the radiation of the air between the measured gas and the non-dispersive infrared detection device, S1 represents the radiation of the air between the background and the non-dispersive infrared detection device, α i (ν) represents the absorption rate of the i-th gas in the gas spectrum database for light with frequency v, C i is the concentration of the gas, and l is the optical path length of light in the gas.

6. The method for analyzing converter flue gas composition based on an AI model according to claim 1, characterized in that: In step S2, the method for constructing a converter flue gas analysis model includes: S21. The concentration data of carbon dioxide gas and carbon monoxide gas are fused with the infrared absorption spectrum curve of flue gas to obtain an AI sub-model; S22. Constructing an oxygen gas concentration calculation model; S23. The hydrogen gas detection concentration is used as an independent coefficient and integrated with the AI ​​sub-model and the oxygen gas concentration calculation model to obtain the converter flue gas analysis model.

7. The method for analyzing converter flue gas composition based on an AI model according to claim 1, characterized in that: In step S3, the non-dispersive infrared detection equipment includes an infrared light source, a non-dispersive infrared sensor and a semiconductor sensor are provided on the opposite side of the infrared light source, an optical filter is provided on the side close to the infrared light source between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor, and a narrow-band filter is provided on the side close to the sensor between the infrared light source and the non-dispersive infrared sensor and the semiconductor sensor.

8. The method for analyzing converter flue gas composition based on an AI model according to claim 1, characterized in that: In step S4, preprocessing includes signal conversion, signal amplification and signal filtering.

Images