CN109852748B - Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate - Google Patents
Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate Download PDFInfo
- Publication number
- CN109852748B CN109852748B CN201910147686.XA CN201910147686A CN109852748B CN 109852748 B CN109852748 B CN 109852748B CN 201910147686 A CN201910147686 A CN 201910147686A CN 109852748 B CN109852748 B CN 109852748B
- Authority
- CN
- China
- Prior art keywords
- gas flow
- distribution
- image
- center
- state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Radiation Pyrometers (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method for monitoring the development process of gas flow in the distribution period of a blast furnace and predicting the utilization rate of gas; the method comprises the following steps: (1) collecting and processing data; (2) processing an infrared image; (3) extracting image features; (4) establishing a dynamic change model of distribution period gas flow distribution; (5) extracting central features of the gas flow; (6) position calibration of the image and the charge level; (7) establishing a dynamic change model of a distribution period gas flow center; (8) establishing a coal gas utilization rate prediction model; the method realizes dynamic tracking of distribution states of gas flows in a distribution period and distribution of gas flow centers at charge level drop points by processing acquired furnace top infrared images and utilizing a clustering algorithm, a statistical method, a feature recognition technology and a pattern recognition technology, and utilizes a neural network algorithm to mine the relation between the development process of the gas flows in the distribution period and the gas utilization rate, thereby realizing real-time prediction of the gas flow utilization rate and providing help for realizing intelligent production.
Description
Technical Field
The invention relates to the field of intelligent production and monitoring in metallurgical industry, in particular to a method for monitoring the development process of gas flow in a blast furnace material distribution period and predicting the utilization rate of gas.
Background
Blast furnace iron making is a typical "black box" production that makes it difficult for operators to understand the dynamics of the gas flow. Each blast furnace has a gas flow development process (including a dynamic change process of gas flow distribution and a dynamic change process of a gas flow center) suitable for the best gas flow development process under a certain raw material condition, and the reasonable gas flow development process of the gas flow of the 'distribution period' is an important condition for ensuring smooth operation of blast furnace burden, normal operation of physicochemical reaction and improvement of the utilization rate of gas. At present, blast furnace operators mainly carry out regulation and control production through blast furnace production information acquired by various sensors and mechanical equipment and long-term accumulated experience, and the regulation and control mode is subject to a large number of subjective factors of the operators and is difficult to realize intelligent production.
In recent years, students establish a mathematical model through data generated by a blast furnace monitoring system, and some students analyze distribution of a charge level temperature field and a relation between the temperature field and a furnace condition by using partial data, which has no statistics, accuracy and universality; still some scholars analyze the relationship between the distribution condition and the utilization rate of the gas flow by taking hours as a unit, and the method cannot reflect the dynamic change of the gas flow and cannot be well applied to the actual blast furnace production. The fact shows that a large amount of data hides important value information behind the data. The core technology of the invention is to carry out statistical analysis on a large amount of infrared video data, track the dynamic change of distribution of gas flow and the dynamic change of a gas flow center in a distribution period in real time by using a statistical method and a pattern recognition technology, and mine the value information between the distribution and the gas utilization rate by using a neural network algorithm, thereby realizing the intelligent monitoring and prediction of blast furnace iron-making.
Disclosure of Invention
The invention aims to solve the problems that the development process of the gas flow in the material distribution period is difficult to track and identify and the relationship between the development process of the gas flow and the utilization rate of the gas is difficult to determine, and provides a method for monitoring the development process of the gas flow in the material distribution period of a blast furnace and intelligently predicting the utilization rate of the gas; the method processes the acquired furnace top infrared image, utilizes a clustering algorithm, a statistical method, a feature recognition technology and a pattern recognition technology to realize dynamic tracking of distribution state of gas flow and distribution of a gas flow center at a charge level drop point, and utilizes a neural network algorithm to dig the relation between the development process of the gas flow in the distribution period and the gas utilization rate, thereby realizing real-time prediction of the gas flow utilization rate and providing help for realizing intelligent production; the method can detect the coal gas flow development process in real time and predict the coal gas flow utilization rate, so that the blast furnace burden distribution operation can realize on-line monitoring, visual control and intelligent prediction.
The invention is realized by the following technical scheme: a method for monitoring the development process of gas flow in the distribution period of a blast furnace and predicting the utilization rate of the gas comprises the following steps:
(1) data acquisition and processing: acquiring three-month blast furnace top infrared video production data, stock rod measurement data and gas utilization rate of each distribution period on line in a certain steel mill, and converting the video into 24-frame per second gray level images through software;
(2) and (3) infrared image processing: carrying out batch superposition processing on the images obtained in the step (1) to obtain 3600 frames of infrared images of the gas flow per hour; removing non-material surface information from the superposed image through gradient interpolation processing; filtering the image obtained by interpolation, and removing noise and pulse interference by adopting a filtering mode combining wavelet transformation and mean filtering; obtaining a relatively clear infrared image;
(3) image feature extraction: solving the corresponding singular value of each frame of image, extracting the first 15% of the singular values to form a vector to represent the image, wherein the vector is called as an image characteristic vector;
(4) establishing a dynamic change model of distribution period gas flow distribution: the specific process is realized through the following steps:
(4.1) introducing the concept of 'blast furnace burden distribution period': the time from the beginning of material distribution to the beginning of the next material distribution is called a blast furnace material distribution period; dividing the blast furnace burden distribution period into a blast furnace burden distribution period and a blast furnace gas flow development period;
(4.2) classifying the images by using the image feature vectors: extracting 3600 multiplied by 24 multiplied by 90 frame image characteristic vectors, namely three-month video data, carrying out cluster analysis by using a bisectionk-means clustering method to obtain different categories and solving clustering center vectors of the categories;
(4.3) a state model of gas flow distribution: calculating the average energy of each image by using the clustering center vectors obtained in the step (4.2), dividing the distribution state of the gas flow according to the average energy of each image, wherein the category with the minimum energy belongs to a first state which is a second state, a third state, a fourth state, a fifth state and a sixth state in sequence;
(4.4) determining the dynamic change characteristic vector of the gas flow distribution: calculating the time construction vector of each state of the gas flow distribution in each distribution period through the established gas flow distribution state model, wherein the vector is called the dynamic change characteristic vector of the gas flow distribution;
(5) extracting the central features of the coal gas flow: because the value ranges of the gray values of the infrared images corresponding to the coal gas flows in different distribution states are different, different threshold values are adopted for different coal gas flow distribution states to carry out global segmentation on the filtered image to obtain a bright zone area where the center of the coal gas flow is located, and the centroid coordinate of the bright zone area is defined as the coordinate of the center of the coal gas flow on the image;
(6) and (3) position calibration of images and charge levels: when the infrared camera shoots a charge level, an inclination angle is formed between the shooting plane and the imaging plane to cause the shot infrared image to have certain inclination deformation, so that the position calibration of pixel points of the image and the charge level is needed to obtain the actual position of the center of the coal gas flow on the charge level, and the actual position is projected onto the furnace throat plane;
(7) establishing a dynamic change model of a distribution period gas flow center: the specific implementation process is as follows:
(7.1) according to the position of the obtained gas flow center on the furnace throat plane, calculating the mass center coordinate of the gas flow center of each state of the gas flow distribution in each distribution period;
(7.2) introducing the concept of 'gas flow center offset degree and offset direction', calculating the offset degree and offset direction of each state gas flow center, and carrying out quantitative processing on the gas flow center:
wherein R is the radius of the furnace throat of the blast furnace,
is the horizontal (vertical) coordinate of the ith state on the material surface under the t-th material distribution period,
the offset degree of the ith state gas flow center at the t period,
the offset direction of the ith state gas flow center under the t-th distribution period utilizes the vector
Representing the central dynamic change characteristic of the gas flow in the tth distribution period;
(8) establishing a coal gas utilization rate prediction model: the coal gas flow distribution dynamic change characteristic vector and the coal gas flow center dynamic change characteristic vector form a new vector in each distribution period
Representing the dynamic process of the development of the coal gas flow in the distribution period; and the relation between the development process of the gas flow in the material distribution period and the corresponding gas utilization rate is excavated by utilizing the radial basis function neural network, so that the real-time prediction of the gas utilization rate is realized, and the intelligent material distribution of the blast furnace is helped.
The invention has the beneficial effects that: (1) tracking the distribution state of the gas flow and the central change of the gas flow in real time through the established dynamic change model of the distribution period gas flow distribution and the dynamic change model of the distribution period gas flow center; (2) the real-time prediction of the gas utilization rate of the distribution period is realized through the relationship between the development process of the gas flow of the distribution period and the gas utilization rate; an optimization algorithm is also provided for the intelligent material distribution of the blast furnace; (3) the method provided by the invention can improve the production efficiency of the blast furnace and realize energy conservation and consumption reduction.
Drawings
FIG. 1 is a blast furnace infrared image (gray scale) obtained in step1 of the example of the present invention.
FIG. 2 is an image obtained after the infrared image processing in step2 according to the embodiment of the present invention.
FIG. 3 is a diagram showing the distribution state of gas flow at each moment of the distribution cycle obtained in step (4.3) of the embodiment of the present invention.
FIG. 4 is a tracing diagram of the gas flow distribution state in a certain period of time in step (4.4) of the embodiment of the present invention.
FIG. 5 is a diagram of the image location of step (5.2) in an embodiment of the present invention.
Fig. 6 is an infrared image pickup plan view of step (6.3) in the embodiment of the present invention.
Fig. 7 is a diagram of the correspondence between the actual shot charge level in step (6.3) and the infrared image in the embodiment of the present invention.
FIG. 8 is a distribution diagram of the positions of the gas flow centers of the development states of the gas flow in a certain distribution period in the throat plane in the step (6.3) of the present invention.
Fig. 9 is a schematic flow diagram of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the present invention will be further described in detail with reference to the following embodiments, which are only used for illustrating the technical solution of the present invention and are not limited.
The invention collects blast furnaces (2500 m) of a certain steel mill3) Production data from 10 to 12 months in 2013 were studied; the method for monitoring the development process of the gas flow in the distribution period of the blast furnace and intelligently predicting the utilization rate of the gas is provided; the method comprises the following steps: (1) collecting and processing data; (2) processing an infrared image; (3) extracting image features; (4) establishing a dynamic change model of distribution period gas flow distribution; (5) extracting central features of the gas flow; (6) position calibration of the image and the charge level; (7) establishing a dynamic change model of a distribution period gas flow center; (8) and establishing a coal gas utilization rate prediction model.
1. Data acquisition and processing
The method mainly comprises the following steps that the data mainly comprise blast furnace top infrared video data within three months, installation parameters and self parameters of an infrared camera, average gas flow utilization rate in each distribution period and trial rod measurement data; as shown in figure 1, the infrared image of 24 frames per second is obtained by processing the online collected blast furnace infrared video through software.
2. Infrared image processing
A series of processing is performed on the obtained infrared image to obtain a more ideal infrared image (as shown in fig. 2), and the specific process is as follows:
(2.1) image superimposition processing
Superposing 24 frames of images in the same second to obtain an infrared image of 1 frame/second; through image superposition processing, the data volume of subsequent processing is reduced, and useful information of the image is not lost;
(2.2) processing of image non-level information
Non-material level information such as infrared image up-sampling time '2013-12-1504: 00: 00' and 'channel three' is subjected to gradient interpolation processing to remove interference information of the image, so that the image information is more accurate;
(2.3) image Filter processing
Blast furnace ironmaking is a highly complex physicochemical reaction process, and infrared images are easily influenced by shooting environments such as noise and pulse interference, so that the infrared images need to be filtered; the infrared image mainly generates Gaussian white noise, the mean filtering processing has good denoising effect, but the inherent defect of the mean filtering exists, and the detail part of the image is damaged while denoising; in order to solve the problem, the invention combines wavelet transformation and mean value filtering to filter the image; therefore, the noise can be reduced, and the detailed part of the image can be kept; the method comprises the following specific steps:
step1: the image containing noise is subjected to one-layer wavelet decomposition by adopting wavelet basis bior1.5 to obtain 3 high-frequency detail subimages
And 1 low frequency detail sub-image
；
Step2: for 3 high-frequency detail sub-images
Respectively carrying out mean value filtering and storing the images after processing
；
Step 3: get Step2 into 3 high frequency detail sub-images
With low-frequency detail sub-images
And reconstructing to obtain a denoised image.
3. Image feature extraction
(ii) the infrared image pixel matrix obtained after the filtering treatment
) Is marked as
(ii) a The singular value corresponding to the matrix hides all information of the image, and the importance of the information and the size of the singular value present positive correlation; thus, a matrix of pixels
Can be decomposed into:
in the formula (1-2), Σ is a diagonal matrix, and its diagonal elements are 288 singular values of X, which are expressed in order as:
；
previous studies have found that the first 15% of the singular values substantially cover the original image information and, therefore, the amount of orientation
To represent the image features at time t, called the image feature vector.
4. Establishing dynamic change model of distribution period gas flow distribution
(4.1) introducing the concept of 'blast furnace burden distribution period': the period from the last material distribution to the next material distribution is called a blast furnace material distribution period; the blast furnace burden distribution period is divided into a blast furnace gas flow development period and a blast furnace burden distribution period (the burden distribution period accounts for about 5% of the gas flow development period, so the method only researches the gas flow development period);
(4.2) classifying the images by using the image feature vectors: clustering method pair by using bisecting k-means
Clustering to obtain different categories (6 categories according to the invention) and calculating the cluster center vector of each category
Each category representing a gas flow distribution profile;
(4.3) a state model of gas flow distribution: knowing that a frame of infrared image corresponds to a matrix X of pixels, the energy E of the matrix X can be represented by a Frobenius norm
And then:
as can be seen from the equation (1-3), the energy of the image can be represented by singular values; thus, the above-obtained cluster center vector
Can represent the average energy characteristics of the ith type of image; the average energy of each type is obtained through the formulas (1-3)
By using
The distribution state of the gas flow is divided, and the first state to which the category with the minimum energy belongs is the second state, the third state, the fourth state, the fifth state and the sixth state in sequence; the distribution state of the gas flow at each moment of the distribution period is obtained by using the method, as shown in figure 3;
(4.4) dynamic tracking of gas flow distribution state: dynamically tracking the distribution state of the gas flow according to the established model, such as a tracking diagram (2013/12/821: 08: 10-2013/12/821: 08: 10) of the distribution state of the gas flow in a certain time period shown in FIG. 4;
table 1 shows the statistical results of the time of each state of seven consecutive fabric cycles in FIG. 4
Unit: second of
(4.5) determining the dynamic change characteristic vector of the gas flow distribution: firstly, determining the distribution state of the gas flow at each moment through the established state model of the distribution of the gas flow, and then counting the time forming vector of each state of the distribution of the gas flow in each distribution period
As shown in table 1; using vectors
Representing the dynamic change characteristics of the gas flow distribution in the t distribution period, and the vector is called the dynamic change characteristic vector of the gas flow distribution.
5. Gas flow center feature extraction
According to the actual production of the blast furnace, the dynamic change of the center of the gas flow can reflect the operation condition of the blast furnace and the utilization rate of the gas flow; according to the blast furnace smelting principle, the higher the temperature of the place where the coal gas flow is stronger, namely a high-temperature area is the area where the center of the coal gas flow is located; as the gray value ranges of the infrared images of the gas flow under different distribution states are different, global threshold segmentation is carried out on the filtering image classification, a bright zone area where the center of the gas flow is located is obtained, and the centroid pixel coordinate of the bright zone is obtained and used as the position of the center of the gas flow on the image; the method comprises the following specific steps:
(5.1) performing classification threshold segmentation on the image: the invention adopts different thresholds to respectively carry out global threshold segmentation on the infrared images of all the distributed states of the gas flow. Setting segmentation threshold values of various images
To obtain a binary image
Wherein
The number of the gas flow development states is represented;
in the formula (5-1)
Is as followsiPixel point under development state
Is determined by the gray-scale value of (a),
is as followsiAn image after development state segmentation;
(5.2) extracting the central features of the gas flow: the size of the known image is
And (3) establishing a rectangular coordinate system (as shown in fig. 5) by taking the center of the image as an origin, and calculating the central coordinate of the bright band area where the center of the gas flow is located, which is obtained in step (5.1), as the coordinate of the center of the gas flow on the infrared image.
6. Position scaling of images and charge levels
When the infrared camera shoots a charge level, an inclination angle is formed between the shooting plane and the imaging plane, so that the infrared image method obtained by shooting has certain inclination deformation, and therefore, position calibration needs to be carried out on each pixel point of the image and the actual charge level to obtain the actual position of the gas flow center on the charge level; the method comprises the following specific steps:
(6.1) establishing a world coordinate system: in order to determine the infrared camera and the position of the dynamically changed charge level, a unified world coordinate system is established; setting the center of a blast furnace throat plane as a coordinate origin (0, 0, 0), taking an intersection line parallel to a shooting plane and an imaging plane in the blast furnace throat plane as an X axis, taking an intersection line vertical to a charge level and the imaging plane in the blast furnace throat plane as a Y axis, and taking a direction vertical to the blast furnace throat plane from top to bottom as a Z axis positive direction;
(6.2) determining the material level position: the stock rod is a device for measuring the charge level change of the blast furnace, the initial position of the stock rod is positioned at the position H above the furnace throat, and the measured value is H; the shape of the blast furnace burden surface has little influence on the infrared image, so the burden surface is assumed to be horizontal in the model established by the invention, and the position of any point Q on the burden surface in the established coordinates is
；
(6.3) image and charge level position calibration: o in FIG. 61Position of infrared camera, O2The camera shoots a focus (the position of the central point of the image on the charge level), O3Is the center of the blast furnace throat charge level, D is the position of the center of the furnace throat charge level on the image, and the included angle of the shooting plane and the imaging plane is
The camera shoots a wide angle of
QH is the throat charge level, KH is the distance between the charge level and the throat plane, O1L and O1B is the installation distance; the corresponding relation between the actual shot charge level and the infrared image is obtained through projection calculation by combining the imaging parameters of the camera (as shown in figure 7); corresponding the gas flow center coordinates in each distribution state of the gas flow obtained in the step (5.2) to the relation to obtain the actual position coordinates of the gas flow center on the charge level, and projecting the coordinates onto the furnace throat plane; FIG. 8 shows the position of the center of the gas flow in the development states of the gas flow on the throat plane in a certain distribution period.
7. Establishing a dynamic change model of a distribution period gas flow center
(7.1) according to the coordinates of the gas flow center on the furnace throat plane obtained in the step (6), the centroid coordinates of the gas flow center in each distribution state of each distribution period are obtained
；
(7.2) introducing the concept of 'integral offset degree and integral offset direction of the gas flow center', calculating the integral offset degree and offset direction of the gas flow center in each state, and carrying out quantitative processing on the center of each distribution state of the gas flow:
in the above formula, R is the radius of the furnace throat of the blast furnace,
is the horizontal (vertical) coordinate of the centroid of the ith state gas flow central point of the tth distribution period on the furnace throat plane,
is the integral offset degree of the gas flow center of the ith state in the tth distribution period,
the integral offset direction of the ith state gas flow center of the tth distribution period is shown; using vectors
Representing the dynamic change characteristics of the gas flow center in the tth distribution period, wherein the vector is called a dynamic change characteristic vector of the gas flow center; table 2 shows the statistical results of the overall positions of the gas flow centers in each distribution state of the gas flow in a certain distribution period;
TABLE 2 statistical results of the overall position of the coal gas flow center in each distribution state of the coal gas flow in a certain distribution period
(7.3) dynamically tracking the gas flow center in the distribution period: by using
And (3) representing the position coordinate of the mass center of the ith state gas flow central point of the tth distribution period on the furnace throat plane, and dynamically tracking the gas flow center according to the coordinate.
8. Establishing a gas utilization rate prediction model
In the actual production of the blast furnace, the coal gas utilization rate is an important mark for reflecting the running state of the blast furnace, and the development dynamic changes of coal gas flows in different distribution periods correspond to different coal gas utilization rates, so that the relationship between the development process of the coal gas flows in the distribution periods and the corresponding coal gas utilization rates is excavated, and the method has important significance for predicting the coal gas utilization rate and realizing intelligent furnace burden.
Using the characteristic vector of the dynamic change of the gas flow distribution in each distribution period (
) And a gas flow center dynamic change feature vector: (
) Form new vectors
Representing the dynamic change characteristic of the development of the gas flow in the t-th distribution period, wherein m is the number of the distribution periods and is called
The dynamic change characteristic vector of the gas flow development in the distribution period is obtained. By using
Predicting corresponding average gas utilization
The method comprises the following specific steps:
step1, let the sample set be
Wherein m is the number of samples; taking 90% of the sample set as a training set, and taking the remaining 10% as a testing set;
step2, normalizing all data by using a formula (8-1), and transferring the data to an interval [0,1 ];
step 3: provided by using MATLAB self-contained neural network tool
The function quickly creates a radial basis function neural network, and the radial basis function neural network is trained through training set data;
and Step4, inputting the test set into the trained neural network model, if the prediction accuracy of the model does not reach a preset value, adding training data until the prediction accuracy of the model reaches the preset value, and storing the network structure for guiding the burden distribution of the blast furnace in the actual blast furnace production.
Description of the drawings: (1) the larger the data sample adopted by the invention is, the more accurate the infrared camera monitoring is, and the more accurate the judgment of the furnace condition of the blast furnace and the prediction of the gas flow are;
(2) the invention can track the development of blast furnace gas flow in real time and give timely blast furnace operation information to blast furnace operators.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that various changes and modifications can be made, and equivalents can be substituted for elements thereof without departing from the scope of the invention.
Claims (1)
1. The method for monitoring the development process of the gas flow in the distribution period of the blast furnace and predicting the utilization rate of the gas comprises the steps of collecting production data of the blast furnace, processing and extracting characteristics of collected furnace top infrared images, utilizing a clustering algorithm, a statistical method, a characteristic identification technology and a pattern identification technology to realize dynamic tracking of the distribution state of the gas flow and the distribution of the gas flow center on a charge level drop point, utilizing a neural network algorithm to mine the relation between the development process of the gas flow in the distribution period and the corresponding utilization rate of the gas, realizing real-time prediction of the utilization rate of the gas flow, and providing help for realizing intelligent distribution, and is characterized in that: the method comprises the following steps:
(1) data acquisition and processing: acquiring three-month blast furnace top infrared video production data, stock rod measurement data and gas utilization rate of each distribution period on line in a certain steel mill, and converting the video into 24-frame per second gray level images through software;
(2) and (3) infrared image processing: carrying out batch superposition processing on the images obtained in the step (1) to obtain 3600 frames of infrared images of the gas flow per hour; removing non-material surface information from the superposed image through gradient interpolation processing; filtering the image obtained by interpolation, and removing noise and pulse interference by adopting a filtering mode combining wavelet transformation and mean filtering; obtaining a relatively clear infrared image;
(3) image feature extraction: solving the corresponding singular value of each frame of image, extracting the first 15% of the singular values to form a vector to represent the image, wherein the vector is called as an image characteristic vector;
(4) establishing a dynamic change model of distribution period gas flow distribution: the specific process is realized through the following steps:
(4.1) introducing the concept of 'blast furnace burden distribution period': the time from the beginning of material distribution to the beginning of the next material distribution is called a blast furnace material distribution period; dividing the blast furnace burden distribution period into a blast furnace burden distribution period and a blast furnace gas flow development period;
(4.2) classifying the images by using the image feature vectors: extracting 3600 multiplied by 24 multiplied by 90 frame image characteristic vectors, namely three-month video data, carrying out cluster analysis by using a bisectionk-means clustering method to obtain different categories and solving clustering center vectors of the categories;
(4.3) a state model of gas flow distribution: calculating the average energy of each image by using the clustering center vectors obtained in the step (4.2), dividing the distribution state of the gas flow according to the average energy of each image, wherein the category with the minimum energy belongs to the first state and sequentially belongs to the second state, the third state, the fourth state, the fifth state and the sixth state;
(4.4) determining the dynamic change characteristic vector of the gas flow distribution: calculating the time construction vector of each state of the gas flow distribution in each distribution period through the established gas flow distribution state model, wherein the vector is called the dynamic change characteristic vector of the gas flow distribution;
(5) extracting the central features of the coal gas flow: because the value ranges of the gray values of the infrared images corresponding to the coal gas flows in different distribution states are different, different threshold values are adopted for different coal gas flow distribution states to carry out global segmentation on the filtered image to obtain a bright zone area where the center of the coal gas flow is located, and the centroid coordinate of the bright zone area is defined as the coordinate of the center of the coal gas flow on the image;
(6) and (3) position calibration of images and charge levels: when the infrared camera shoots a charge level, an inclination angle is formed between the shooting plane and the imaging plane to cause the shot infrared image to have certain inclination deformation, so that the position calibration of pixel points of the image and the charge level is needed to obtain the actual position of the center of the coal gas flow on the charge level, and the actual position is projected onto the furnace throat plane;
(7) establishing a dynamic change model of a distribution period gas flow center: the specific implementation process is as follows:
(7.1) according to the position of the obtained gas flow center on the furnace throat plane, calculating the mass center coordinate of the gas flow center of each state of the gas flow distribution in each distribution period;
(7.2) introducing the concept of 'gas flow center offset degree and offset direction', calculating the offset degree and offset direction of each state gas flow center, and carrying out quantitative processing on the gas flow center:
wherein R is the radius of the furnace throat of the blast furnace,
is the horizontal (vertical) coordinate of the ith state on the material surface under the t-th material distribution period,
the offset degree of the ith state gas flow center at the t period,
the offset direction of the ith state gas flow center under the t-th distribution period utilizes the vector
Representing the central dynamic change characteristic of the gas flow in the tth distribution period;
(8) coal for buildingGas utilization rate prediction model: the coal gas flow distribution dynamic change characteristic vector and the coal gas flow center dynamic change characteristic vector form a new vector in each distribution period
Representing the dynamic process of the development of the coal gas flow in the distribution period; and the relation between the development process of the gas flow in the material distribution period and the corresponding gas utilization rate is excavated by utilizing the radial basis function neural network, so that the real-time prediction of the gas utilization rate is realized, and the intelligent material distribution of the blast furnace is helped.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910147686.XA CN109852748B (en) | 2019-02-27 | 2019-02-27 | Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910147686.XA CN109852748B (en) | 2019-02-27 | 2019-02-27 | Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109852748A CN109852748A (en) | 2019-06-07 |
| CN109852748B true CN109852748B (en) | 2020-05-08 |
Family
ID=66899331
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910147686.XA Active CN109852748B (en) | 2019-02-27 | 2019-02-27 | Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109852748B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111593151B (en) * | 2020-04-30 | 2022-03-22 | 柳州钢铁股份有限公司 | On-line detection method for depth of blast furnace tap hole |
| CN112063782A (en) * | 2020-08-31 | 2020-12-11 | 北京科技大学 | Blast furnace burden surface clustering method |
| CN112946232A (en) * | 2021-02-04 | 2021-06-11 | 成都秦川物联网科技股份有限公司 | Natural gas energy metering data acquisition method and system |
| CN113034480B (en) * | 2021-04-01 | 2023-12-19 | 艾德领客(上海)数字技术有限公司 | Blast furnace damage analysis method based on artificial intelligence and image processing |
| CN114998326B (en) * | 2022-07-19 | 2023-03-07 | 江苏瑞立环保工程股份有限公司 | Method and system for detecting distribution condition of gas flow based on artificial intelligence |
| CN115541030B (en) * | 2022-09-21 | 2024-02-27 | 中冶南方工程技术有限公司 | Method and device for identifying blast furnace top charge level temperature distribution and storage medium |
| CN115578612B (en) * | 2022-10-11 | 2023-08-04 | 浙江大学 | Blast furnace top distribution stage identification method and device based on marker target detection |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11140515A (en) * | 1997-11-07 | 1999-05-25 | Sumitomo Metal Ind Ltd | Operation of blast furnace |
| CN103409569A (en) * | 2013-08-29 | 2013-11-27 | 武汉钢铁(集团)公司 | Method for controlling airflow distribution by utilizing furnace top infrared image |
| CN105002321A (en) * | 2015-06-16 | 2015-10-28 | 内蒙古科技大学 | Method for processing coal gas flow center dynamic tracking and monitoring coal gas utilization rate |
| CN107679581A (en) * | 2017-10-23 | 2018-02-09 | 内蒙古科技大学 | The method of characteristic value gas treatment flow distribution based on infrared image picture element matrix |
-
2019
- 2019-02-27 CN CN201910147686.XA patent/CN109852748B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11140515A (en) * | 1997-11-07 | 1999-05-25 | Sumitomo Metal Ind Ltd | Operation of blast furnace |
| CN103409569A (en) * | 2013-08-29 | 2013-11-27 | 武汉钢铁(集团)公司 | Method for controlling airflow distribution by utilizing furnace top infrared image |
| CN105002321A (en) * | 2015-06-16 | 2015-10-28 | 内蒙古科技大学 | Method for processing coal gas flow center dynamic tracking and monitoring coal gas utilization rate |
| CN107679581A (en) * | 2017-10-23 | 2018-02-09 | 内蒙古科技大学 | The method of characteristic value gas treatment flow distribution based on infrared image picture element matrix |
Non-Patent Citations (1)
| Title |
|---|
| 基于大数据下煤气流不同发展时期温度场的定标研究;张生海 等;《内蒙古科技大学学报》;20181231;第37卷(第4期);第343-348页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109852748A (en) | 2019-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109852748B (en) | Method for monitoring development process of gas flow in distribution period of blast furnace and predicting gas utilization rate | |
| CN105002321B (en) | A kind of gas treatment stream center dynamic tracking and gas utilization rate monitoring method | |
| CN103886612B (en) | Automatic water level extraction method and system based on reservoir monitoring camera | |
| CN107478657A (en) | Stainless steel surfaces defect inspection method based on machine vision | |
| CN110428450B (en) | Scale-adaptive target tracking method applied to mine tunnel mobile inspection image | |
| CN103324951B (en) | A kind of polynary isomery monitoring water environment data evaluation and method for early warning | |
| CN108694349B (en) | Pantograph image extraction method and device based on linear array camera | |
| CN101751603A (en) | Online bar section image automatic counting device and method | |
| CN109993154A (en) | The lithium sulfur type instrument intelligent identification Method of substation's simple pointer formula | |
| CN112862744B (en) | Intelligent detection method for internal defects of capacitor based on ultrasonic image | |
| CN115797354A (en) | Method for detecting appearance defects of laser welding seam | |
| CN106529441B (en) | Depth motion figure Human bodys' response method based on smeared out boundary fragment | |
| CN116229052B (en) | Method for detecting state change of substation equipment based on twin network | |
| CN107230219A (en) | A kind of target person in monocular robot is found and follower method | |
| CN103578121B (en) | Method for testing motion based on shared Gauss model under disturbed motion environment | |
| CN112614139B (en) | Conveyor belt ore briquette screening method based on depth map | |
| CN107025439B (en) | Lip region feature extraction and normalization method based on depth data | |
| CN110956178B (en) | Plant growth measuring method and system based on image similarity calculation and electronic equipment | |
| CN112784922A (en) | Extraction and classification method of intelligent cloud medical images | |
| CN112288765A (en) | Image processing method for vehicle-mounted infrared pedestrian detection and tracking | |
| CN111582076A (en) | Picture freezing detection method based on pixel motion intelligent perception | |
| CN115546799A (en) | Backlight-free water meter liquid crystal display screen display number identification method under poor lighting condition | |
| CN107832723A (en) | Smog recognition methods and system based on LBP gaussian pyramids | |
| Zhao et al. | MSC-AD: A Multiscene Unsupervised Anomaly Detection Dataset for Small Defect Detection of Casting Surface | |
| CN105426811A (en) | Crowd abnormal behavior and crowd density recognition method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |