A method, device and system for temperature detection in the iron supply process section

Abstract

This invention discloses a method, apparatus, and system for temperature detection in the iron feeding process section. The method for temperature detection in the iron feeding process section includes: determining the temperature at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point; determining the temperature at any one of the skimmer, trough, swing nozzle, or ladle temperature detection points includes: acquiring a thermal image and emissivity of the temperature detection point; determining an emissivity correction coefficient based on the thermal image; correcting the emissivity using the emissivity correction coefficient to obtain a corrected emissivity; and determining the temperature of the temperature detection point using the corrected emissivity.

Description

Technical Field

[0001] The embodiments of the present invention relate to iron feeding process technology, and more particularly to a method, device and system for temperature detection in the iron feeding process section. Background Technology

[0002] Blast furnace ironmaking is a key process in the steelmaking process. High-quality molten iron is the prerequisite for smelting "high-purity" high-quality steel and the foundation for improving the quality of steel products.

[0003] Temperature monitoring of molten iron is a key step in molten iron quality control. Generally, the temperature of molten iron during tapping is affected by the following factors: heat dissipation during the tapping process (the amount of heat dissipation is affected by the size of the iron flow, the length of the iron channel, the iron feeding time, and the temperature of the ladle before the iron is fed); heat dissipation during the process of adding scrap steel to the ladle (the time from when the ladle is finished tapping to when the steelmaking process begins); and heat dissipation during the time when the ladle is waiting to be put into the furnace for steelmaking.

[0004] Currently, temperature detection during tapping is mainly achieved through the following methods: hot metal samples are taken manually or using a robotic arm from the main trough for analysis to obtain the composition and temperature of the hot metal; or the temperature is measured manually or using a portable thermocouple at the main trough location. When using manual sampling, intermittent sampling is employed. After tapping from the blast furnace, 2-3 samples are typically taken randomly, with sampling intervals generally between 30-50 minutes.

[0005] Based on the above, in the existing technology, it is difficult to continuously detect the temperature of molten iron during the iron receiving process, and the trend of molten iron energy change during the tapping process of the blast furnace cannot be accurately grasped, which is not conducive to better guiding the blast furnace process operation. Summary of the Invention

[0006] This invention provides a method, device, and system for detecting temperature in the iron feeding process section, so as to achieve continuous detection of molten iron temperature and improve the accuracy of temperature measurement.

[0007] In a first aspect, embodiments of the present invention provide a method for detecting the temperature of the iron supply process section, comprising:

[0008] Determine the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point.

[0009] Determining the temperature of any one of the following temperature detection points: skimmer temperature detection point, trough temperature detection point, swing nozzle temperature detection point, or ladle temperature detection point, includes:

[0010] Acquire thermal images and emissivity of the temperature detection points;

[0011] Determine the emissivity correction factor based on the thermal image;

[0012] The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity;

[0013] The temperature at the temperature detection point is determined by the corrected emissivity.

[0014] Optionally, acquiring thermal images of the temperature detection points includes:

[0015] The original thermal image of the temperature detection point is acquired, and the occlusion and interference elements in the original thermal image are eliminated to obtain the thermal image.

[0016] Optionally, determining the emissivity correction factor based on the thermal image includes:

[0017] Using the thermal image as input, the first model is used to extract the image features of the thermal image to obtain the image feature vector;

[0018] The image feature vector is input, and the emissivity correction coefficient is determined using the second model.

[0019] Optionally, the thermal image is input, and the first model is used to extract the image features of the thermal image to obtain an image feature vector, including:

[0020] A sequence of thermal images is obtained, and the image features of each thermal image are extracted sequentially using the first model to generate an image feature sequence.

[0021] The image feature vector is input, and the emissivity correction coefficient is determined using the second model, including:

[0022] Using the image feature sequence, the correction coefficients for each image feature are determined sequentially by the second model to generate a correction coefficient sequence;

[0023] Also includes:

[0024] The emissivity correction coefficients are determined using a third model based on the correction coefficient sequence.

[0025] Optionally, the first model includes a convolutional neural network model, and the second model includes a recurrent neural network model.

[0026] Optionally, the third model includes a support vector regression model.

[0027] Optionally, determining the temperature at any temperature detection point includes:

[0028] Acquire thermal images of at least two bands at the temperature detection point, and obtain the emissivity of the temperature detection point;

[0029] Determine the single-band correction coefficient corresponding to the thermal image of each band, and use all the single-band correction coefficients to determine the emissivity correction coefficient;

[0030] The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity;

[0031] The temperature at the temperature detection point is determined by the corrected emissivity.

[0032] Optionally, the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point are used as the basis for adjusting the flow rate of molten iron in the iron feeding process section.

[0033] Secondly, embodiments of the present invention also provide a temperature detection device for the iron feeding process section, including a temperature detection unit, the temperature detection unit being used for:

[0034] Determine the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point.

[0035] Determining the temperature of any one of the following temperature detection points: skimmer temperature detection point, trough temperature detection point, swing nozzle temperature detection point, or ladle temperature detection point, includes:

[0036] Acquire thermal images and emissivity of the temperature detection points;

[0037] Determine the emissivity correction factor based on the thermal image;

[0038] The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity coefficient;

[0039] The temperature at the temperature detection point is determined by the corrected emission coefficient.

[0040] Thirdly, embodiments of the present invention also provide a temperature detection system for the iron feeding process section, including a controller, wherein the controller is configured with an executable program, and the executable program is used to implement the temperature detection method for the iron feeding process section described in the embodiments of the present invention when it is executed.

[0041] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention proposes a temperature detection method for the iron feeding process section. In this method, thermal images and emissivity of selected temperature detection points are acquired. Based on a neural network model, an emissivity correction coefficient is determined through the thermal image. The emissivity is then corrected using the correction coefficient, and the temperature of the temperature detection point is determined using the emissivity. This enables automated temperature detection. Based on the emissivity correction coefficient determined through the neural network model, the influence of environmental factors on temperature measurement accuracy can be reduced, thus improving the accuracy of temperature measurement. Furthermore, the selected temperature detection points in this method include the slag skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point. These temperature detection points cover the main nodes of the iron feeding process section, enabling continuous temperature detection throughout the entire iron tapping process. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the iron-feeding process section in the embodiment;

[0043] Figure 2 This is a flowchart of the temperature detection method in the embodiment;

[0044] Figure 3 This is a flowchart of another temperature detection method in the embodiment;

[0045] Figure 4 This is a flowchart of another temperature detection method in the embodiments;

[0046] Figure 5 This is a flowchart of another temperature detection method in the embodiments;

[0047] Figure 6 This is a schematic diagram of the temperature detection system for the iron supply process section in the embodiment. Detailed Implementation

[0048] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0049] Example 1

[0050] Figure 1 This is a schematic diagram of the iron-feeding process section in the embodiment, for reference. Figure 1 In this embodiment, the iron supply process section mainly includes a main channel section and a secondary channel section. A skimmer is installed in the main channel section, and a swing nozzle is installed in the secondary channel section, with iron ladles installed on both sides of the swing nozzle.

[0051] During smelting operations, the high-temperature slag and iron in the blast furnace flow out of the taphole and first pass through the main ditch. The slag and iron mixture is separated inside the main ditch due to the large density difference. Under the obstruction of the slag skimmer, the upper layer of slag flows into the slag ditch and is water-quenched to form water slag.

[0052] The denser molten iron flows into the trough along the lower channel of the lower skimmer, then passes through the swing nozzle (which can tilt and swing left and right depending on the position of the ladle), and then flows into the ladle. Once one side of the ladle is filled with the expected weight of molten iron, the swing nozzle is moved to the other side of the ladle to complete the continuous iron feeding process.

[0053] In this embodiment, a temperature detection point is set at the slag skimmer, the trough, the swing nozzle, and the ladle. By determining the temperature of each of the above temperature detection points (i.e., the temperature of the slag skimmer, the trough, the swing nozzle, and the ladle), the temperature of the molten iron flowing through the iron feeding process section can be continuously detected.

[0054] For example, in this embodiment, the temperature detection points for the skimmer, the trough, the swing nozzle, and the ladle can be determined in the following manner:

[0055] The point with the lowest dust concentration in the areas of the skimmer, the trough, the nozzle, and the ladle is determined respectively. These points are then used as the temperature detection points for the skimmer, the trough, the nozzle, and the ladle, respectively.

[0056] For example, in a preferred embodiment, the temperature detection points for the skimmer, the trough, the swing nozzle, and the ladle are determined in the following manner:

[0057] Starting from the through hole after the skimmer and ending at the highest permissible liquid level inside the ladle, set N temporary temperature detection points at even intervals along the path of the molten iron flow, denoted as N1 to N2. n ;

[0058] The temperature at each temporary temperature detection point was measured multiple times manually using a portable thermometer, and the average temperature at each temporary temperature detection point was obtained and denoted as T1 to T2. n ;

[0059] Using average temperature T1 and average temperature T n Determine the maximum temperature loss during molten iron flow, and denote the maximum temperature loss as T. s Then T s =T n -T1;

[0060] The position of the through hole after the skimmer is used as the temperature detection point of the skimmer.

[0061] Starting from the average temperature T1, determine that the difference between the average temperature T1 and the average temperature T2 is equal to T. s / 4 Average temperature T x1 , will be related to the average temperature T x1 The corresponding temporary temperature detection point will serve as the temperature detection point for the Longgou ditch.

[0062] Determine the average temperature T x1 The difference is equal to T s / 4 Average temperature T x2 , will be related to the average temperature T x2 The corresponding temporary temperature detection point is used as the nozzle temperature detection point;

[0063] The difference between the average temperature T2 and the temperature T is determined to be equal to T. s / 4 Average temperature T x3 , will be related to the average temperature T x3 The corresponding temporary temperature detection point serves as the temperature detection point for the iron ladle.

[0064] For example, in the above scheme, if the difference between the average temperature of the last temporary temperature detection point located in the Longgou section and the average temperature T1 is greater than T... s / 4, then adjust the structure of the dragon ditch, including:

[0065] During the downtime for ditch repair, the inclination angle of the ditch near the skimmer end was increased by adjusting the thickness of the molten iron layer. A horizontal groove was added at the very end of the ditch (the purpose of adding the horizontal groove is to reduce the excessive impact force on the nozzle caused by excessively fast molten iron flow, while ensuring that the angle between the start and end points of the ditch remains unchanged). This reduced the length of the inclined section near the skimmer end of the ditch until the average temperature at the last temporary temperature monitoring point of the ditch section was approximately equal to T1. s / 4;

[0066] If the inclination angle increases, the horizontal groove will be longer. If the inclination angle increases by 1 degree, the length of the horizontal groove will increase by 100 mm.

[0067] For example, in the above scheme, if the structure of the trough is adjusted, the temperature detection points of the trough, the temperature detection points of the swivel nozzle, and the temperature detection points of the iron ladle need to be redefined.

[0068] For example, in this embodiment, the same method is used to determine the temperature of each of the above temperature detection points. Figure 2 This is a flowchart of the temperature detection method in the embodiment, for reference. Figure 2 Determining the temperature of any one of the following temperature detection points: the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, or the ladle temperature detection point, includes:

[0069] S101. Acquire thermal images and emissivity of the temperature detection point.

[0070] For example, in this embodiment, a spectral thermometer is provided at the temperature detection point, and the image generated by the spectral thermometer is used as the thermal image of the temperature detection point.

[0071] For example, in this embodiment, the method of obtaining emissivity is not specifically limited. For example, the emissivity of the temperature detection point can be determined by methods such as reflectivity method or multi-wavelength method.

[0072] S102. Determine the emissivity correction factor based on the thermal image.

[0073] For example, in this embodiment, the emissivity correction coefficient can be determined by means of a neural network model. When using a neural network model, the thermal image is used as input and the emissivity correction coefficient is used as output (or one of the output vectors of the neural network model).

[0074] For example, in this embodiment, the type and structure of the neural network model used are not specifically limited. For example, neural network models such as BP, CNN, and RNN can be used.

[0075] S103. The emissivity is corrected using an emissivity correction factor to obtain the corrected emissivity.

[0076] For example, in this embodiment, the product of the emissivity correction factor and the emissivity can be used as the corrected emissivity.

[0077] S104. Determine the temperature of the temperature detection point by correcting the emissivity.

[0078] For example, in this embodiment, a mathematical model representing the relationship between emissivity and temperature can be obtained, and the temperature of the temperature detection point can be determined based on the mathematical model.

[0079] For example, the mathematical model described above can be determined by function fitting using experimental measurement data or publicly available data in the literature. In this embodiment, the method of function fitting is not limited, and the fitted mathematical model can be a linear model, an exponential model, or other forms of function model.

[0080] For example, in this embodiment, in addition to continuously detecting the temperature of the molten iron flowing through the iron feeding process section, the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point can also be used as the basis for adjusting the flow rate of molten iron in the iron feeding process section.

[0081] For example, in one possible implementation, the flow rate of molten iron can be adjusted as follows:

[0082] Determine the temperature difference between two adjacent temperature detection points among the following: the temperature of the skimmer, the temperature of the trough, the temperature of the swing nozzle, and the temperature of the ladle.

[0083] If any temperature difference exceeds the set value, the opening diameter of the iron tap will be increased to increase the flow rate of molten iron, while the suction capacity (power) of the dust removal system will be reduced to achieve stable molten iron energy.

[0084] If any temperature difference is less than the set value, the opening diameter of the iron tap will be reduced to reduce the flow rate of molten iron, while the suction capacity (power) of the dust removal system will be increased to achieve stable molten iron energy.

[0085] For example, the above setting value can be set based on experience; for instance, the setting value can be T. s / 4.

[0086] For example, in this solution, the control of molten iron flow rate based on the temperature of the temperature detection point can accurately guide blast furnace operation, which can play a certain role in stabilizing molten iron quality and reducing production quality accidents.

[0087] This embodiment proposes a temperature detection method for the iron feeding process section. The method acquires thermal images and emissivity of selected temperature detection points. Based on a neural network model, an emissivity correction coefficient is determined from the thermal images. The emissivity is then corrected using this coefficient, and the temperature at the detection point is determined based on the emissivity. This enables automated temperature detection. The emissivity correction coefficient determined by the neural network model reduces the impact of environmental factors on temperature measurement accuracy, thus improving accuracy. Furthermore, the selected temperature detection points include those at the skimmer, the trough, the swing nozzle, and the ladle. These points cover the main nodes of the iron feeding process section, enabling continuous temperature detection throughout the entire iron feeding process.

[0088] Figure 3 This is a flowchart of another temperature detection method in the embodiment, see reference. Figure 3 As one possible implementation, the method can also be:

[0089] S201. Obtain the emissivity of the temperature detection point.

[0090] S202. Obtain the original thermal image of the temperature detection point, and eliminate occlusion and interference elements in the original thermal image to obtain the thermal image.

[0091] For example, in this solution, the image generated by the spectrometer is used as the original thermal image of the temperature detection point.

[0092] For example, in this solution, eliminating occlusion interference elements in the original thermal image mainly involves eliminating the problem of abnormal grayscale values ​​of certain pixels caused by dust occlusion in the original thermal image.

[0093] For example, in this solution, occlusion interference elements in the original thermal image can be eliminated to obtain the thermal image in the following way:

[0094] Generate an original image that is the same size as the original thermal image (same number of pixels, same pixel arrangement);

[0095] Acquire multiple consecutive frames of raw thermal images. For each pixel in the raw thermal image, determine whether there are abnormal changes in the gray level of the same pixel in the raw thermal image sequence (e.g., in a certain frame of raw thermal image, the gray level of the same pixel suddenly increases or decreases significantly).

[0096] If the gray level of a pixel changes abnormally, it is determined that there are occlusion interference elements in the original thermal image. At this time, the gray level of the pixel in the original image is fixed to the gray level value of the pixel when the gray level is normal in a frame of the original thermal image.

[0097] If the gray level of a pixel does not change abnormally, then the gray level of that pixel in the original image is fixed to the gray level value of that pixel in one frame of the original thermal image.

[0098] Once the grayscale values ​​of all pixels in the original image are determined, the original image becomes a thermal image.

[0099] For example, if there is dust interference when the spectrometer is performing measurements, the gray level of the corresponding pixel in the original thermal image is usually within a fixed gray level range. When the gray level of a certain pixel changes abnormally, the gray level value of that pixel in the original thermal image sequence when the gray level is normal can be determined based on the above gray level range.

[0100] For example, the method for determining the thermal image in this step can also be applied to Figure 1 In the scheme shown, replace the corresponding content in step S101.

[0101] S203. Using the thermal image as input, the first model is used to extract the image features of the thermal image to obtain the image feature vector.

[0102] S204. Input the image feature vector and use the second model to determine the emissivity correction coefficient.

[0103] Combining steps S203 and S204, in this scheme, the emissivity correction coefficient is determined using the first model and the second model.

[0104] For example, in this scheme, the first model adopts a convolutional neural network (CNN) model, and the second model adopts a neural network (NN) model.

[0105] For example, in this solution, the CNN model structure is set according to requirements. For example, the CNN can be specifically ResNet-18, ResNet-50, etc.

[0106] For example, in this scheme, the model structure of the NN is related to the image feature vector and the set model output, wherein the model output of the NN is set to be at least a one-dimensional vector, which represents the emissivity correction coefficient.

[0107] S205. The emissivity is corrected using an emissivity correction factor to obtain the corrected emissivity.

[0108] S206. Determine the temperature of the temperature detection point by correcting the emissivity.

[0109] In this plan, Figure 1 Based on the beneficial effects of the proposed scheme, this scheme preprocesses the image generated by the spectral thermometer to eliminate abnormal pixels caused by dust obstruction, thereby improving the calculation accuracy of the emissivity correction coefficient.

[0110] Figure 4 This is a flowchart of another temperature detection method in the embodiments, see reference. Figure 4 ,exist Figure 3 Based on the scheme shown, this method can also be:

[0111] S301. Obtain the emissivity of the temperature detection point.

[0112] S302. Obtain the original thermal image of the temperature detection point, and eliminate occlusion and interference elements in the original thermal image to obtain the thermal image.

[0113] S303. Obtain the thermal image sequence, and use the first model to extract the image features of each thermal image in sequence to generate the image feature sequence.

[0114] For example, in this scheme, the first model adopts the ResNet-18 model, and the input of ResNet-18 is a thermal image, and the output is a 512-dimensional image feature vector.

[0115] In this scheme, a specified number of consecutive thermal images (e.g., 100 frames) are acquired, and the image feature vector of each thermal image is determined by the ResNet-18 model. The set of all image feature vectors is then used as the image feature sequence.

[0116] S304. Using the image feature sequence, the correction coefficients of each image feature are determined sequentially through the second model to generate a correction coefficient sequence.

[0117] For example, in this scheme, the second model adopts a recurrent neural network (RNN) model. The input of the RNN is set to an image feature vector, and the output is a two-dimensional vector. The two elements in the two-dimensional vector are set to the temperature measurement quality coefficient and the correction coefficient, respectively.

[0118] For example, in the above scheme, the temperature measurement quality coefficient is used to represent the accuracy of determining the temperature using the corresponding frame of thermal image.

[0119] For example, in this scheme, the set of all two-dimensional vectors is used as the sequence of correction coefficients.

[0120] S305. Based on the correction coefficient sequence, the emissivity correction coefficients are determined using the third model.

[0121] For example, in this scheme, the third model adopts the Support Vector Regression (SVR) model. The input of the SVR model is set to a high-dimensional matrix formed by the sequence of correction coefficients, and the output is the emissivity correction coefficient (or the emissivity correction coefficient category, where one emissivity correction coefficient corresponds to one emissivity correction coefficient).

[0122] S306. The emissivity is corrected by using an emissivity correction factor to obtain the corrected emissivity.

[0123] S307. Determine the temperature of the temperature detection point by correcting the emissivity.

[0124] For example, in this solution, there are no improvements to the structure of ResNet-18, RNN, and SVR models, and the sample data used to train each model can be obtained through calibration experiments.

[0125] exist Figure 2 Based on the beneficial effects of the scheme shown, this scheme determines the correction coefficient sequence based on the thermal image sequence, and then determines the emissivity correction coefficient based on the third model. This can avoid the problem of inaccurate calculation of the emissivity correction coefficient due to accidental factors when determining the emissivity correction coefficient through a single frame thermal image.

[0126] Figure 5 This is a flowchart of another temperature detection method in the embodiments, see reference. Figure 5 As one possible implementation, the method can also be:

[0127] S401. Obtain the emissivity of the temperature detection point.

[0128] S402. Acquire thermal images of at least two bands at the temperature detection point.

[0129] For example, in this solution, the temperature detection point can be configured with a multi-channel spectrometer, and the image of one channel of the multi-channel spectrometer can be used as a thermal image of one band.

[0130] S403. Determine the single-band correction coefficient corresponding to the thermal image of each band, and use all the single-band correction coefficients to determine the emissivity correction coefficient.

[0131] For example, in this step, assuming there are N bands, determining the emissivity correction coefficient specifically includes:

[0132] The thermal image sequence for each band is obtained separately. For each thermal image sequence, the first model is used to extract the image features of each thermal image in turn to generate an image feature sequence.

[0133] For example, in this scheme, a thermal image sequence is defined as containing 100 frames of thermal images, and the image feature sequence of the i-th band is denoted as {x}. i,1 ,…,x i,100}

[0134] For example, in this scheme, the first model uses the ResNet-18 model, and the image features x are set. i,j It is a 512-dimensional image feature vector.

[0135] For each band of image feature sequence, the correction coefficients of the image feature sequence are determined sequentially by the second model, and a correction coefficient sequence is generated.

[0136] For example, in this scheme, the second model adopts an RNN model, and the input of the RNN is set to a 512-dimensional image feature x. i,j The output is a two-dimensional vector y. i,j In this context, the two elements of the two-dimensional vector are defined as the temperature measurement quality coefficient and the correction coefficient, respectively.

[0137] For example, in this scheme, the output of the RNN model includes {y i,1 ,…,y i,100}, y i,100 As the correction coefficients for the image feature sequence of the i-th band, the correction coefficient sequence is {y}. 1,100 ,…,y N,100}

[0138] Based on the correction coefficient sequence, the emissivity correction coefficients are determined using the third model.

[0139] For example, in this scheme, the third model adopts the SVR model, and the input of the SVR model is set to a high-dimensional matrix formed by the sequence of correction coefficients, and the output is the emissivity correction coefficient.

[0140] S404. The emissivity is corrected by using an emissivity correction factor to obtain the corrected emissivity.

[0141] S405. Determine the temperature of the temperature detection point by correcting the emissivity.

[0142] exist Figure 1 Building upon the beneficial effects of the previously presented scheme, this scheme acquires a multi-band thermal image sequence and determines the emissivity correction coefficient based on the multi-band thermal images. This avoids the problem of large uncertainty in the calculated emissivity correction coefficient due to the unknown emissivity coefficient on the surface of the temperature being measured being a non-ideal blackbody. Furthermore, this scheme employs a CNN model with strong image feature extraction capabilities to determine image feature vectors, and an RNN model with strong feature extraction capabilities to determine the correction coefficient sequence. Using the correction coefficient sequence, the emissivity correction coefficient is then determined through an SVR model, which improves the execution efficiency of the method and the accuracy of the calculated emissivity correction coefficient.

[0143] Example 2

[0144] This embodiment proposes a temperature detection device for the iron feeding process section, including a temperature detection unit, which is used for:

[0145] Determine the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point.

[0146] Determining the temperature of any one of the following temperature detection points: skimmer temperature detection point, trough temperature detection point, swing nozzle temperature detection point, or ladle temperature detection point, includes:

[0147] Acquire thermal images and emissivity of the temperature detection points;

[0148] Determine the emissivity correction factor based on the thermal image;

[0149] The emissivity is corrected using an emissivity correction factor to obtain the corrected emissivity factor;

[0150] The temperature at the temperature detection point is determined by correcting the emission coefficient.

[0151] In this embodiment, the specific implementation method and beneficial effects of the temperature detection unit are as follows: Figure 2 The proposed solution is the same, and the specific details will not be repeated here.

[0152] For example, in this embodiment, the temperature detection unit can also be configured to use... Figure 3 , Figure 4 or Figure 5 The scheme shown achieves temperature detection, and its specific implementation method and beneficial effects will not be elaborated further.

[0153] Example 3

[0154] Figure 6 This is a schematic diagram of the temperature detection system in the iron feeding process section of the embodiment, for reference. Figure 6 The system may include: a controller 100 and several multi-channel spectral thermometers (1 to n).

[0155] For example, in this embodiment, the multi-channel spectral thermometer is respectively installed at the temperature detection points of the upper skimmer, the trough, the swing nozzle, and the ladle. Specifically, refer to... Figure 1 A temperature detection point can be set at the slag skimmer and the ditch, two temperature detection points can be set symmetrically at the swing nozzle, and one temperature detection point can be set at each iron ladle position.

[0156] For example, in this embodiment, the controller 100 is configured with an executable program, which is used to implement any of the temperature detection methods for the iron feeding process section described in Embodiment 1 when the executable program is run.

[0157] refer to Figure 6 In one possible implementation, the system may further include a server 200, a data storage device 300, and a local area network device 400.

[0158] The multi-channel spectral thermometer is connected to the controller 100 via server 200. Server 200 is also connected to data storage device 300 and local area network device 400.

[0159] For example, in this solution, the controller 100, server 200, data storage device 300, local area network device 400, and multi-channel spectral thermometer constitute a distributed system;

[0160] The controller 100 can be located in the central control room, and the server 200, data storage device 300, and local area network device 400 can be located in the area covered by the local area network.

[0161] For example, in this solution, a data storage device 300 is configured to store historical temperature measurement records; the local area network device 400 can be an office computer, etc.

[0162] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for detecting temperature in the iron feeding process section, characterized in that, include: Determine the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point. The position of the through hole after the skimmer is used as the temperature detection point of the skimmer. Self-averaging temperature Starting from, determine the average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point will serve as the temperature detection point for the Longgou ditch. Determine and average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point is used as the nozzle temperature detection point; Determine and average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point is used as the ladle temperature detection point; Determining the temperature of any one of the following temperature detection points: skimmer temperature detection point, trough temperature detection point, swing nozzle temperature detection point, or ladle temperature detection point, includes: Acquire thermal images and emissivity of the temperature detection points; Determine the emissivity correction factor based on the thermal image; The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity; The temperature at the temperature detection point is determined by the corrected emissivity. Determining the emissivity correction factor based on the thermal image includes: Using the thermal image as input, the first model is used to extract image features from the thermal image to obtain an image feature vector, including: A sequence of thermal images is obtained, and the image features of each thermal image are extracted sequentially using the first model to generate an image feature sequence. Using the image feature vector as input, the second model is used to determine the emissivity correction coefficient, including: Using the image feature sequence, the correction coefficients for each image feature are determined sequentially by the second model to generate a correction coefficient sequence; Also includes: The emissivity correction coefficients are determined using a third model based on the correction coefficient sequence. The first model includes a convolutional neural network model, and the second model includes a recurrent neural network model; The third model includes the support vector regression model.

2. The method for detecting temperature in the iron supply process section as described in claim 1, characterized in that, Acquiring thermal images of temperature detection points includes: The original thermal image of the temperature detection point is acquired, and the occlusion and interference elements in the original thermal image are eliminated to obtain the thermal image.

3. The method for detecting temperature in the iron supply process section as described in claim 1, characterized in that, Determining the temperature at any temperature detection point includes: Acquire thermal images of at least two bands at the temperature detection point, and obtain the emissivity of the temperature detection point; Determine the single-band correction coefficient corresponding to the thermal image of each band, and use all the single-band correction coefficients to determine the emissivity correction coefficient; The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity; The temperature at the temperature detection point is determined by the corrected emissivity.

4. The method for detecting temperature in the iron supply process section as described in claim 1, characterized in that, The temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point serve as the basis for adjusting the flow rate of molten iron in the iron feeding process section.

5. A temperature detection device for an iron feeding process section, characterized in that, Includes a temperature detection unit, which is used for: Determine the temperatures at the skimmer temperature detection point, the trough temperature detection point, the swing nozzle temperature detection point, and the ladle temperature detection point. The position of the through hole after the skimmer is used as the temperature detection point of the skimmer. Self-averaging temperature Starting from, determine the average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point will serve as the temperature detection point for the Longgou ditch. Determine and average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point is used as the nozzle temperature detection point; Determine and average temperature The difference equals average temperature , will be with average temperature The corresponding temporary temperature detection point is used as the ladle temperature detection point; Determining the temperature of any one of the following temperature detection points: skimmer temperature detection point, trough temperature detection point, swing nozzle temperature detection point, or ladle temperature detection point, includes: Acquire thermal images and emissivity of the temperature detection points; Determine the emissivity correction factor based on the thermal image; The emissivity is corrected using the emissivity correction coefficient to obtain the corrected emissivity coefficient; The temperature at the temperature detection point is determined by the corrected emission coefficient. Determining the emissivity correction factor based on the thermal image includes: Using the thermal image as input, the first model is used to extract image features from the thermal image to obtain an image feature vector, including: A sequence of thermal images is obtained, and the image features of each thermal image are extracted sequentially using the first model to generate an image feature sequence. Using the image feature vector as input, the second model is used to determine the emissivity correction coefficient, including: Using the image feature sequence, the correction coefficients for each image feature are determined sequentially by the second model to generate a correction coefficient sequence; Also includes: The emissivity correction coefficients are determined using a third model based on the correction coefficient sequence. The first model includes a convolutional neural network model, and the second model includes a recurrent neural network model; The third model includes the support vector regression model.

6. A temperature detection system for an iron feeding process section, characterized in that, The system includes a controller configured with an executable program, which, when executed, is used to implement the temperature detection method for the iron feeding process section as described in any one of claims 1 to 4.

Images