WO2021197168A1 - Image recognition-based automatic deslagging method and system using converter slag remaining process - Google Patents

Abstract

Disclosed in the present invention is an image recognition-based automatic deslagging method using a converter slag remaining process, comprising the steps of: determining a target deslagging amount on the basis of the weight of slag generated during converter smelting and a target amount of remained slag during smelting using a slag remaining process; acquiring and storing a real-time image of a hot-state slag stream during converter deslagging; determining characteristic boundaries of the hot-state slag stream; selecting a tracking area; according to image recognition, calculating, in real time, the volume Vi of slag in the tracking area and the total weight Gt of the slag poured out at a deslagging time t; when the volume Vi of the slag in the tracking area is greater than a set measurement threshold Vmin, controlling, on the basis of a weight change rate dGt/dt, a converter to tilt in real time at an angular velocity ωi, and obtaining a real-time tilting angle of the converter; calculating a real-time position of a slag ladle trolley, so that a slag ladle receives the slag poured out from the converter in real time; and when a difference value between a total deslagging weight Gt and a target deslagging amount G0 is a preset deviation value, stopping the tilting of the converter and starting the shaking-back of the converter. Also disclosed in the present invention is an image recognition-based automatic deslagging system using a converter slag remaining process.

Description

Method and system for automatic slag dumping of converter slag retention process based on image recognition Technical field

The invention relates to a method and system for automatic slag dumping, in particular to a method and system for automatic slag dumping in a converter slag retention process.

Background technique

At present, in the world, the converter smelting of major steel plants basically adopts model steelmaking, which realizes the one-key operation of the blowing process, and as the converter automatic tapping technology has been put into use, the tapping process of the converter is also changed. The way of manual operation is changed to automatic operation. In converter smelting, there is a so-called "converter slag retention process" which is widely used in converter steelmaking technology. The process utilizes iron oxide and calcium oxide from the final slag of the previous furnace to reduce the consumption of iron and steel materials and auxiliary materials. At the same time, it is beneficial to the formation of the early slag of the next furnace and helps the early dephosphorization. The characteristic of this process is that not all the slag is poured out when dumping the slag, but a certain amount of slag is reserved for the next smelting.

The converter slag dumping that usually adopts the slag retention process includes the following steps:

1) Drive the slag tank trolley (also called the slag bag trolley) to the position of receiving slag;

2) The converter is tilted, and the slag in the converter is poured out from the mouth of the large furnace;

3) Operate the slag pot trolley so that the slag is poured into the slag pot on the slag pot trolley;

4) Meet the requirement of the target slag retention of the steel grade "slag retention process", and stop the slag dumping;

5) Shake the converter back to the vertical position to complete the entire slag dumping process.

It should be pointed out that since the viscosity of the slag is related to the characteristics of steel smelting and the slagging process, according to the viscosity of the slag, the slag is generally divided into four types: A, B, C, and D. The viscosity of the slag is different, the speed of the slag flowing out and the location of the drop point are also different. Therefore, in the slag dumping process of the existing converter, it is currently relying on manual operation of the converter tilting, while observing the falling point of the slag flow to move the position of the slag tank trolley to ensure that the converter tilting and the slag tank trolley are completed. Accurate matching.

However, such operations have high technical requirements for the operators, and there is a risk that the slag will pour out of the slag pot due to improper operations by the operators, which may cause safety accidents. In addition, smoke and splashing will be generated during the slag dumping process, which has higher requirements for personnel safety protection.

The publication number CN103397134A, the publication date is November 20, 2013, and the Chinese patent document titled "A method for calculating the amount of slag retained according to the tilt angle" discloses a method of calculating the amount of slag retained. This method is to tilt the converter after the end of blowing. When the slag in the converter just flows out of the furnace mouth, record the tilt angle of the converter at this time. The tilt angle is the critical angle ɑ, which can be calculated by the critical angle ɑ The total volume of molten steel in the furnace is measured by the sub-lance to obtain the volume of molten steel in the converter. The tilting angle β of the converter after the slag pour is recorded to obtain the total volume of the remaining molten body, and finally combined with the lining erosion data provided by the lining measurement system Calculate the total volume of the melt in the furnace before the slag dumping and the volume of the dumped slag. The difference between the two is the volume of the slag remaining in the converter. According to the density of slag, the amount of slag remaining can be calculated. This method does not require additional monitoring and weighing equipment.

In addition, the publication number is CN107502698A, the publication date is December 22, 2017, and the Chinese patent document entitled "An automated steelmaking method suitable for slag-less smelting" discloses a calculation model for slag dumping and slag retention. Slag dumping model: on the basis of the original automatic steelmaking model, the heat and oxygen balance calculation of the early slag is added, and the target alkalinity of the early stage is added. Based on the on-site smelting data and laboratory simulation, the best time interval for the early slag dumping will be added. In the early-stage blowing control of the model, different blowing times are calculated according to the different components of the molten iron. The establishment of the slag retention model is to add a slag weight weighing device on the slag truck to collect the weight of the slag dumped in the early stage and the end, and transfer the data to the automated steelmaking model through the first stage.

It should be noted that although the above-mentioned prior art discloses models for slag dumping and slag retention, the purpose of the model is only to improve the converter slag retention process, and does not involve how the converter slag retention process realizes automatic slag dumping. Although there is a converter automatic tapping technology in the prior art, it does not involve the converter fully automatic slag dumping technology. Automatic converter slagging is to pour the hot slag of the converter from the large furnace mouth of the converter, and the automatic converter tapping is to pour molten steel from the tapping mouth of the converter. Although both are automatic tilting converters and the trolley runs automatically, there are similarities, but the two technologies are essentially different. Therefore, it is necessary to adopt new technology to control and realize the automatic slag dumping of the converter slagging process based on image recognition.

Summary of the invention

One of the objectives of the present invention is to provide a method for automatic slag dumping of the converter slag retention process based on image recognition, which realizes the automatic slag dumping of the converter slag retention process without weighing detection of the slag tank. Image recognition technology is used to determine the start of slag dumping and the characteristic boundary of the slag flow image, and the rate of change of the slag weight obtained by the slag volume model based on image recognition can be used to control the angular velocity of the automatic tilting of the converter, according to the real-time tilting angle of the converter The position model of the slag tank trolley can realize the accurate matching of the automatic tilting of the converter and the automatic traveling of the slag tank trolley. The method for automatic slag dumping of the converter slag retention process based on image recognition of the present invention is simple, accurate, safe and reliable. It can not only replace manual operations, reduce the labor intensity of related operators, but also reduce the weight of the slag tank. The cost of equipment transformation has a wide range of applicability.

In order to achieve the above objective, the present invention proposes an automatic slag dumping method based on image recognition for the converter slag retention process, which includes:

Based on the slag weight G ₁ produced in the converter smelting process and the target slag remaining amount G _{2 in} smelting using the slag-retaining process, determine the target slag dumping amount G ₀ =G ₁ -G ₂ ;

Collect and store real-time images of the hot slag flow during the converter slag dumping process;

Determine the characteristic boundary of the hot slag flow based on image recognition;

Based on the characteristic boundary of the hot slag flow, select the tracking area in the path of the hot slag flow falling from the furnace mouth to the slag pot;

_{Calculate the volume of slag V i} at each moment in the tracking area _{and the total weight G t of} slag dumping when the slag dumping time is t based on image recognition;

When the tracking of the slag in the region of the volume V _i> after setting the detection threshold value V _min, image recognition, calculated on the total weight of the deslagging G _t, based on the weight change rate dG _t / _dt, real-time control of the converter tilt angle velocity ω _i , and obtain the real-time tilting angle of the converter according to the real-time tilting angular velocity, calculate the real-time position of the slag tank trolley based on the real-time tilting angle of the converter, and drive the slag tank trolley to the target position by controlling the motor. Make the slag pot on the slag pot trolley receive the slag poured from the converter in real time;

When the difference between the total weight of dumped slag G _t and the target amount of dumped slag G ₀ is the preset deviation amount, stop the converter tilting and start the converter rollback.

Here, the difference between the total weight of slag dumping G _t and the target slag dumping amount G ₀ is the preset deviation amount, which means that the total weight of dumping slag G _t can be greater than, less than or equal to the target slag dumping amount G ₀ . In production, the preset deviation depends on the accuracy requirements of each factory, and the preset deviation is generally within the range of ±1t～±3t.

It should be noted that the "slag tank" mentioned in this technical solution can also be referred to as a "slag bag", and the concepts of the two are interchangeable.

In the above technical solution of the present invention, the method for automatic slag dumping based on image recognition in the converter slagging process of the present invention realizes that the slag dumping can be judged through image recognition without the need for weighing detection of the slag tank. The beginning and the characteristic boundary of the slag flow image, use the rate of change of the slag weight obtained by the slag volume model based on image recognition to control the angular velocity of the automatic tilting of the converter. According to the real-time tilting angle and the position model of the slag tank trolley, It can realize the automatic running of the slag tank trolley, and realize the accurate matching of the automatic tilting of the converter and the automatic running of the slag tank trolley.

Further, in the method for automatic slag dumping of a converter slagging process based on image recognition of the present invention, in the step of collecting and storing a real-time image of the hot slag flow during the converter slag dumping process, the real-time image includes : One of visible light image, infrared image and far-infrared image.

Further, in the method based on image recognition process remaining slag is converter slag automatic reverse of the present invention, in the image recognition calculation based on the tracking area within each time the slag volume V _i and the time t deslagging In the step of the total weight of dumped slag G _t , the following content is used to calculate the volume of slag V _i at each moment in the tracking area and the total weight of dumped slag G _t when the dumping time is t based on image recognition:

Divide the tracking area into i×j grids, scan the tracking area line by line with a scan width Δh, and the intersection of the feature boundary and the grid in each scan width forms a micro area;

Calculate the slag thickness based on the following model:

Where δ _j represents the thickness of the slag within a scan width at any time;

Represents the average width of the slag within a sweep width at any time; α represents the coefficient related to the slag viscosity, the greater the slag viscosity, the greater α (which is dimensionless), and its value range is 0<α≤1; λ is The coefficient (it is dimensionless), and its value range is 0<λ≤1; n is the coefficient (it is dimensionless), and its value range is 0＜n≤2;

Then in the micro area with the sweep width of △h, the volume of the slag is

Sum the volume of each micro area to obtain the volume V _i of the slag in the tracking area at the current moment;

When it is detected that the volume of slag in the tracking area _Vi > the set detection threshold V _min , the time at this time is marked as 0, and when the slag dumping time reaches t, the total weight G _{t of the} slag dumping is based on the following The formula of the model is obtained:

Where ρ _slag represents the density of the slag, and g represents the acceleration due to gravity.

In the above scheme, it should be noted that g represents the acceleration of gravity. The acceleration of gravity is affected by many factors in actual operation. It can be 9.8 meters per square second (m/s ² ).

Further, in the method for automatic slag dumping based on image recognition in the converter slagging process of the present invention:

When the volume of slag in the tracking area _Vi ≤ the set detection threshold V _min , control the tilting angular velocity of the converter to the set first angular velocity ω ₁ , at this time ω _i =ω ₁ ;

When the volume of slag in the tracking area _Vi > the set detection threshold V _min , control the tilting angular velocity of the converter to the set second angular velocity ω ₂ , and ω ₂ <ω ₁ , at this time ω _i = ω ₂ ;

When the volume of slag in the tracking area _Vi > the set detection threshold V _{min , and the total weight G t} of the dumped slag calculated based on image recognition satisfies

And when |G _t -G ₀ |>G _min , control the real-time tilting angular velocity of the converter

Where t represents the converter slag dumping time; G _min represents the allowable deviation of the slag weight, which is the set value; k ₁ represents the change rate of the slag weight; b represents the adjustment coefficient, and its value range is 0＜b≤1; The total weight of slag G _t satisfies

And when |G _t -G ₀ |>G _min , control the real-time tilting angular velocity of the converter ω _i =0;

Further, in the method for automatic slag dumping of a converter slag retention process based on image recognition of the present invention, in the above steps, the real-time position x _{i of the} slag tank trolley is controlled according to the following model:

x _i = x ₀ + βθ _i

Among them, x ₀ represents the initial position of the slag tank trolley; β represents the adjustment coefficient, which ranges from 0.01 to 0.1, and the unit parameter is meters/degree (m/°); θ _i = θ ₀ +ω _i t, Where θ _i represents the real-time tilting angle of the _{converter, θ 0} represents the initial tilting angle of the converter, t represents the slag _{dumping time of the converter, and ω i} is the real-time tilting angular velocity of the converter.

In the above scheme, the real-time position x _i of the slag tank trolley can be determined according to the model, and the real-time travel displacement (x _i -x ₀ ) of the slag tank trolley is sent to the control system that controls the travel of the slag tank trolley, and the motor drives the slag tank. The tank trolley travels to the target position x _i determined according to the model, so that the slag tank can accurately receive the hot slag flow.

When |G _t -G ₀ |≤G _min , that is: within the control range of the target weight, the converter stops tilting; starts automatic reversing, and according to the set reversing model, the converter performs reversing until the converter returns to vertical Bit, the automatic slag dumping is over.

Correspondingly, another object of the present invention is to provide an automatic slag dumping system for a converter slag retention process based on image recognition. The automatic slag dumping system for a converter slag retention process based on image recognition is used for slag dumping, in addition to replacing the existing technology In addition to the manual slag dumping operation method, the equipment transformation cost brought by the weighing of the slag tank is reduced.

In order to achieve the above objective, the present invention proposes an automatic slag dumping system for converter slag retention technology based on image recognition, which includes:

Converter

The converter tilting driving device is connected with the converter to drive the converter to tilt and perform the slag dumping action;

Converter tilt angle measuring device, used to measure the tilt angle of the converter;

The slag tank is set on the slag tank trolley;

The slag tank trolley position detection device is used to detect the real-time position of the slag tank trolley;

Image acquisition device for acquiring real-time images of the hot slag flow during the slag dumping process of the converter;

An image processing module for processing the real-time image;

The storage module is used to store the real-time image and/or the real-time image processed by the image processing module;

The image recognition and calculation module is used to recognize the real-time image to determine the characteristic boundary of the hot slag flow, and to calculate the volume of the slag in the selected tracking area in real time;

The control module is respectively connected with the converter tilting drive device, the converter tilt angle measuring device, the slag tank trolley, the slag tank trolley position detection device, the image recognition and calculation module, and the control module is set to: calculate based on image recognition when deslagging is t region of each track within a selected time slag volume V _i and the total weight G deslagging time _t; slag when the trace in the region of the volume V _i> set detection threshold V _min, Based on the total weight G _{t of the} dumped slag, the real-time tilting angular velocity ω _{i of the} converter is controlled, and the real-time tilting angle of the converter is obtained according to the real-time tilting angular velocity, and the real-time position of the slag tank trolley is calculated based on the real-time tilting angle of the converter , So that the slag tank can accept the slag poured from the converter in real time; when _{the difference between the total weight of the slag dumped G t} and the target slag _{dumping amount G 0} is the preset deviation amount, stop the converter tilting and start the converter rollback.

Further, in the automatic slag dumping system of the converter slag retention process based on image recognition of the present invention, it further includes a production process computer system connected to the control module, and the production process computer system sends converter smelting data to the control module .

In the above solution, the automatic slag dumping system for converter slag retention based on image recognition of the present invention further includes a production process computer system, which is connected to the control module and can send converter smelting data to the control module. Corresponding smelting information can be stored in the production process computer system, and the relevant information in the production process computer system can be calculated by using calculation methods commonly used in the prior art in the art to obtain the total slag weight, and obtain the target slag residue in the slag retention process In order to obtain the target amount of slag dumping.

Further, in the automatic slag dumping system of the converter slag retention process based on image recognition of the present invention, the image acquisition device includes one of a visible light camera, an infrared camera, and a far infrared camera.

Further, in the automatic slag dumping system for the converter slag retention process based on image recognition of the present invention, the converter inclination angle measuring device includes an encoder or an inclinometer.

Further, in the automatic slag dumping system for the converter slag retention process based on image recognition of the present invention, the slag tank trolley position detection device includes a non-contact distance meter.

Compared with the prior art, the method and system for automatic slag dumping of the converter slagging process based on image recognition according to the present invention has the following advantages and beneficial effects:

The method for automatic slag dumping of the converter slag retention process based on image recognition of the present invention can utilize a set of image acquisition device, and realizes the automatic dumping of the converter slag retention process based on image recognition technology under the condition of a slag-free tank weighing device. Scum. The method for automatic slag dumping of converter slag retention technology based on image recognition is simple and accurate, effectively reducing equipment investment and maintenance costs, not only can replace the manual operation method of the prior art, reduce the labor intensity of related operators, and improve the working environment , Also has a very wide range of applicability, can be applied to various production lines.

In addition, the automatic slag dumping system of the converter slag retention process based on image recognition of the present invention also has the above-mentioned advantages and beneficial effects.

Description of the drawings

FIG. 1 is a schematic structural diagram of a system for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application;

2 is a schematic flow chart 1 of a method for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application;

3 is a schematic diagram of the operation flow of the system for automatic slag dumping of converter slag retention process based on image recognition provided by the embodiment of the application;

FIG. 4 is a second schematic flow diagram of a method for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application;

FIG. 5 is a slag feature boundary recognition diagram of a method for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application.

Detailed ways

Hereinafter, the method and system for automatic slag dumping of the converter slagging process based on image recognition of the present invention will be further described according to the specific embodiments of the present invention and the accompanying drawings, but the description does not constitute an improper limitation of the present invention.

FIG. 1 is a schematic structural diagram of a system for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application.

As shown in FIG. 1, in this embodiment, the converter is connected to the converter tilting drive device, and the converter tilting drive device drives the converter to tilt to perform the slag dumping action.

When the converter starts the slag dumping operation, the converter is tilted so that the converter can be tilted from the vertical position to the slag dumping position at a certain angle to the ground. At this time, the tilting angle of the converter can be measured by the converter inclination angle measuring device. Among them, the converter The inclination measuring device may include an encoder or an inclinometer.

When the converter is dumped to the slag dumping position, the converter starts the slag dumping operation. The converter slag in the converter can flow from the large furnace mouth of the converter to the slag pot (also called slag trolley) on the slag pot trolley (also called slag trolley). Is the slag bag). Among them, the image acquisition device can collect real-time images of the hot slag flow of the converter during the slag dumping process, and can transmit the collected real-time image data to a computer system. The computer system can include: a storage module, an image processing module, and image recognition And computing module, control module. In addition, the automatic slag dumping system for the converter slag retention process based on image recognition of the present invention may include a slag tank trolley position detection device, and the slag tank trolley position detection device may include a non-contact distance meter. The position detection device of the slag tank trolley can be connected with the control module in the computer system to realize real-time monitoring of the position of the slag tank trolley, and detect the real-time position of the slag tank trolley, so that the slag tank is always aligned and receives hot slag flow.

In addition, the automatic slag dumping system of the converter slag retention process based on image recognition of the present invention also includes a production process computer system. The production process computer system can be connected to the control module in the computer system and can send data to the control module. Converter smelting data.

It should be noted that the image processing module can process the real-time image of the hot slag flow, convert the real-time image of the hot slag flow into a grayscale image, and can binarize the grayscale image. The storage module can store the real-time image of the hot slag flow transmitted by the image acquisition device and the real-time image processed by the image processing module. The image recognition and calculation module can recognize the real-time image, determine the characteristic boundary of the hot slag flow, and calculate the volume of the slag in the selected tracking area and the weight of the poured slag in real time. The control module can be connected to the converter, the converter inclination angle measurement device, the slag tank trolley, the slag tank trolley position detection device, the image recognition and calculation module, and the control module can be based on the data transmitted by the converter inclination angle measurement device and the image recognition of the slag dumping process. The data and the slag weight change data calculated based on the image recognition calculation model to adjust the angular velocity of the converter's automatic tilting, so as to realize the converter's automatic tilting. The slag pot platform can be calculated according to the position model of the slag pot trolley. The real-time position of the slag tank trolley is input to the control system that controls the movement of the slag tank trolley. The motor of the slag tank trolley drives the slag tank trolley to travel to the target position determined by the model to make the slag tank The tank trolley can always bear the hot slag flow dumped by the converter.

It should also be noted that because the temperature and color of the hot slag stream are very different from the temperature and color of the surrounding environment, the image acquisition device can be a visible light camera, an infrared camera, or a far-infrared camera. Collect real-time images of the hot slag flow during the converter slag dumping process, and the storage module stores the collected real-time images. The storage module can store the files continuously recorded by the image acquisition device in digital format to facilitate subsequent processing.

FIG. 2 is a schematic flow chart 1 of a method for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application.

As shown in FIG. 2, in this embodiment, the method for automatic slag dumping of the converter slag retention process based on image recognition may include the following contents:

_{Calculate the weight of slag G 1} produced in the converter smelting process through related smelting information such as the amount of auxiliary materials added, the amount of molten iron, the composition of the molten iron, and the target composition in the computer system of the production process;

_{Obtain the target slag retention G 2 in the} smelting process using the slag retention process from the computer system of the production process;

Determine the target dumping amount G ₀ ; G ₀ =G ₁ -G ₂ ;

Collect and store real-time images of the hot slag flow during the converter slag dumping process;

Using the image recognition method, determine the characteristic boundary of the hot slag flow, and identify and calculate the average slag width of the slag surface in each scan width line;

Here, in the method of image recognition, the characteristic boundary of the hot slag flow is determined, and the average slag width of the slag surface in each scan width line is identified and calculated, in the path of the slag flow falling from the converter furnace mouth to the slag pot. Select a tracking area, determine the characteristic boundary of the hot slag flow in the tracking area according to the image recognition method, and identify and calculate the average slag width of the slag surface in each scan width row;

According to the model, calculate the total weight G _{t of the} slag dumping when the slag dumping time is t;

_{Here, in the total weight G t} of the slag dumping when the slag dumping time is t calculated according to the model, the real-time image of the tracking area can be scanned line by pixel, and the volume and weight of the slag in the tracking area at the current time can be calculated according to the model. ; inverted every time the tracking process in the slag weight in the region of the slag summing integrator, can be obtained deslagging time when the total weight of the deslagging _t G t;

According to the calculated change model of the slag volume and weight, the angular velocity of the converter's automatic tilting is controlled to realize the converter's automatic tilting;

According to the real-time tilting angle of the converter and the position model of the slag tank trolley, the real-time position of the slag tank trolley is automatically calculated, and the real-time travel displacement of the slag tank trolley is input to the control system that controls the slag tank trolley. The motor of the tank trolley drives the slag tank trolley to the target position determined according to the model, so that the slag tank can receive the hot slag flow dumped by the converter;

Calculate the weight of the poured slag G _{t in} real time and compare it with the target dumped slag amount G _0. When |G _t -G ₀ |≤G _min , that is, within the control range of the target weight, the converter stops tilting;

Here, it should be noted that G _t represents the weight of the poured slag, G ₀ represents the target slag volume, G _min represents the preset deviation of the slag weight control, G _t , G ₀ and G _min, and G ₁ and G ₂ The unit parameters of these weights can all be kilograms (kg).

After the converter stops tilting, it will start the automatic reversing. According to the set reversing model, the converter will re-shake until the converter returns to the vertical position, and the automatic slag dumping ends.

FIG. 3 is a schematic diagram of the operation flow of the system for automatic slag dumping of the converter slag retention process based on image recognition according to an embodiment of the application.

As shown in Figure 3, in this embodiment, the converter position is in the vertical position, confirm that the slag tank trolley has been cleared at the starting position (ie 0), and the displayed weight of the slag tank trolley is cleared (the weight is 0 kg). ), the converter terminates the blowing, and the equipment and public auxiliary conditions of the related system (referring to the normal state of the equipment hydraulic pressure, lubricating oil, etc.) meet the slag dumping conditions. Click the "automatic slag dumping start" button, the converter will automatically tilt according to the tilting model. The image acquisition device installed at the slag dumping site can record the video of the hot slag flow in real time and transmit the signal to the storage module of the computer system. Perform digital continuous storage.

In this embodiment, it can be implemented according to the following operating steps:

During the converter slag dumping operation, press the converter "automatic slag dumping start" button, the converter will automatically tilt, and the tilt angle of the converter will automatically tilt according to the model of formula (1):

That is θ _i =θ ₀ +ω _i t(1), where θ _i represents the real-time tilting angle of the _{converter, θ 0} represents the initial tilting angle of the converter, and the unit can be degrees (°), and t represents the converter slag dumping time , The unit can be seconds (s), ω _i represents the real-time tilting angular velocity of the converter, and the unit can be degrees/second (°/s).

At this time, the converter _{starts to tilt at the first angular velocity ω 1} , and the converter tilt angle is θ ₁ =ω ₁ t. At this time, ω _i =ω ₁ , the volume of slag in the tracking area _Vi ≤ the set detection threshold V _min . It should be noted that in order to improve the efficiency of slag _{dumping and reduce the slag dumping time, the speed of ω 1} is generally higher, and the specific size is determined according to different converter structures, characteristics of tilting motors, and operating habits. The weight of the slag poured out at this time G _t =0;

When the volume of the slag within the tracking area V _i> set the detection threshold value V _min, it illustrates the flow of hot slag flows into the region of the tracking range. At this time, the control converter tilt angle to the second angular velocity [omega] ₂ is set, i.e. the converter tilting angular velocity ω _i = ω _2, ω ₂ may be a specific size depending on the structure of the converter tilting motor characteristics and operating practices determined, And ω ₂ <ω ₁ . Among them, V _min represents the set detection threshold, and the technician can select and set the corresponding V _min according to the needs in the actual application.

Then based on the image recognition calculation every time the tracking area of the slag volume V _i and deslagging time is t, the total weight of the slag is poured G _t, G _t is calculated on the total weight of the slag can be poured with reference to FIG.

The weight G _{t of the} poured slag satisfies:

And |G _t -G ₀ |>G _min .

The real-time tilting speed of the converter

Among them, in formula (2), t represents the converter slag dumping time, and the unit can be seconds (s); k ₁ represents the set rate of change of slag weight, and the unit can be kilograms per second (kg/s); b can represent The dimensionless adjustment coefficient has a value range of 0<b≤1, and b can be adjusted according to the specific converter equipment structure and actual production conditions. The unit parameters of ω _i , the first angular velocity ω ₁ and the second angular velocity ω ₂ can all be degrees/second (°/s).

The total weight G _{t of the} poured slag satisfies:

And |G _t -G ₀ |>G _min , the converter stops at the current position, and ω _i =0.

In addition, as the tilt angle of the converter changes, the real-time position x _{i of the} slag tank trolley can be controlled based on the model of formula (3):

x _i = x ₀ + βθ _i (3), where x ₀ represents the initial position of the slag tank trolley; β represents the adjustment coefficient, which ranges from 0.01 to 0.1, and the unit parameter can be m/degree (m/ °); θ _i represents the real-time tilting angle of the converter.

According to the real-time position x _i of the slag tank trolley determined by the model, the real-time travel displacement (x _i -x ₀ ) of the slag tank trolley is sent to the control system that controls the slag tank trolley, and the motor drives the slag tank trolley to move. To the target position x _i determined according to the model, the slag tank can accurately receive the hot slag flow.

When the weight of the slag satisfies the condition of the formula |G _t -G ₀ |≤G _min , that is, when _{the difference between the weight G t of the} discharged slag and the target amount of slag G ₀ is calculated as the preset deviation amount, the pour is stopped. Slag, the signal is sent to the motor of the converter tilting, and the converter starts to swing back to the vertical position. The rollback can be controlled according to the formula (4) model: θ _i = θ _max -ω _i t(4), where θ _max represents the converter angle at the end of slag dumping, and the unit parameters of _{θ max} and θ _i and θ _{0 are both} It can be degrees (°), ω _i represents the angular velocity of the converter during reversing, and t represents the time for raising the furnace, and the unit can be seconds (s); θ _i represents the real-time tilting angle of the converter.

It should be noted that those skilled in the art set the value of the _{converter tilting angular velocity ω i} according to the converter capacity, structure and operation process of each factory. The inclination angle of the converter can be rotated freely within the range of 360°. The definition of the positive or negative inclination angle or the inclination position is different according to the operating habits of each production line, but its setting does not affect the setting of the function model, that is to say, the applicability of the function model Wide, can be effectively applied to various production lines.

FIG. 4 is a schematic diagram of the second flow chart of a method for automatic slag dumping of a converter slagging process based on image recognition according to an embodiment of the application.

As shown in Figure 4, the weight of the poured slag calculated based on image recognition can be implemented as follows:

Select the tracking area;

Here, you can use the box to select the tracking area, and divide the tracking area into i×j grids.

Identify feature boundaries;

Here, since the temperature and color of the hot slag flow are very different from the temperature and color of the surrounding environment, the boundary position between the slag flow and the surrounding environment can be used as the characteristic boundary, and the hot slag can be identified according to the image recognition method. The characteristic boundary of the flow. It should be noted that there are many existing image recognition technologies in this field, which can be used to identify the characteristic boundary of the hot slag flow.

Calculate the average width of the hot slag flow on the image.

Here, the tracking area can be scanned line by line with a scan width Δh, and the intersection of the feature boundary and the i×j grid in each scan width forms a micro area. This micro-zone is the hot slag area, and the average width of the hot slag flow in the micro-zone is used

Express.

Calculate the thickness of the slag.

Here, in general, the wider the width of the hot slag stream, the thicker the corresponding thickness of the hot slag stream, and the slag with higher viscosity is thicker than the slag with lower viscosity. The model of formula (5) can be used to calculate the slag thickness:

Among them, δ _j represents the thickness of the slag within a scan width at any time;

Represents the average width of the slag within a sweep width at any time; α represents the coefficient related to the slag viscosity, the greater the slag viscosity, the greater the α, the greater α (which is dimensionless), and the value range is 0<α≤ 1; λ is a coefficient (it is dimensionless), and its value range is 0<λ≤1; n is a coefficient (it is dimensionless), and its value range is 0＜n≤2;

Calculate the volume of slag in the tracking area.

Here, formula (6) can be used to first calculate the volume of slag in the micro-zone with a sweep width of △h:

Then, by summing the volume of each micro area, the volume V _i of the slag in the tracking area at the current time can be obtained.

Calculate the weight of the slag poured at any time.

Here, when it is detected that the volume of slag in the tracking area _Vi > the set detection threshold V _min , the time at this time is marked as 0, that is, t=0.

Then when the dumping time reaches t, the total weight of dumping G _t can be obtained by the model of formula (7):

Among them, ρ _slag represents the density of the slag, and its unit parameter can be kilograms/cubic meter (kg/m ³ ), which can be obtained from public literature according to the main composition of the slag, or obtained through testing. g represents the acceleration of gravity, and g can be It is 9.8 meters per square second (m/s ² ).

FIG. 5 is a slag feature boundary recognition diagram of a method for automatic slag dumping of a converter slag retention process based on image recognition according to an embodiment of the application.

As shown in Figure 5, when identifying the slag feature boundary, you can use a box to select the tracking area and divide this tracking area into i×j grids. The two black solid lines in Figure 5 schematically show the thermal state The characteristic boundary of the slag flow. Among them, the intersection of the characteristic boundary of the hot slag flow and the i×j grid within each scan width of the tracking area will form a micro area ABCD, which is the hot slag area.

In summary, it can be seen that the automatic slag dumping method of the converter slag retention process based on image recognition of the present invention can adopt a set of image acquisition device to realize the converter slag retention process without weighing and detecting the slag tank. Fully automatic slag dumping. The image recognition technology is used to judge the start of slag dumping and the characteristic boundary of the hot slag flow image, and the rate of change of the slag weight obtained by the slag volume model can be used to control the angular velocity of the converter's automatic tilting, according to the real-time tilting angle of the converter and the slag The position model of the running of the tank trolley can accurately match the automatic tilting of the converter and the automatic running of the trolley. The method for automatic slag dumping of the converter slag retention process based on image recognition of the present invention is simple, accurate, safe and reliable. It can not only replace manual operations, reduce the labor intensity of related operators, but also can effectively avoid multiple image acquisition systems. Difficulties caused by installation and maintenance, and reduced the cost of equipment transformation brought by the weighing of the slag tank, and has a wide range of applicability.

In addition, the automatic slag dumping system of the converter slag retention process based on image recognition of the present invention also has the above-mentioned advantages and beneficial effects.

It should be noted that the prior art part of the protection scope of the present invention is not limited to the embodiments given in this application document, and all prior art that does not contradict the solution of the present invention includes but is not limited to the previous Patent documents, prior publications, prior publications, etc. can all be included in the protection scope of the present invention.

In addition, the combination of various technical features in this case is not limited to the combination described in the claims of this case or the combination described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless Contradictions arise between each other.

It should also be noted that the above-listed embodiments are only specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and the subsequent similar changes or modifications are those skilled in the art can directly derive or easily associate from the disclosure of the present invention, and they should all fall within the protection scope of the present invention. .

Claims (10)

1. A method for automatic slag dumping of a converter slag retention process based on image recognition, which is characterized in that it includes:

   Based on the slag weight G ₁ produced in the converter smelting process and the target slag remaining amount G _{2 in} smelting using the slag-retaining process, determine the target slag dumping amount G ₀ =G ₁ -G ₂ ;

   Collect and store real-time images of the hot slag flow during the converter slag dumping process;

   Determine the characteristic boundary of the hot slag flow based on image recognition;

   Based on the characteristic boundary of the hot slag flow, select the tracking area in the path of the hot slag flow falling from the furnace mouth to the slag pot;

   _{Calculate the volume of slag V i} at each moment in the tracking area _{and the total weight G t of} slag dumping when the slag dumping time is t based on image recognition;

   When the slag volume V _{i in} the tracking area> the set detection threshold V _min , based on the total weight of dumped slag G _t , the real-time tilting angular velocity ω _{i of the} converter is controlled, and the converter is obtained according to the real-time tilting angular velocity The real-time tilting angle of the converter is calculated based on the real-time tilting angle of the converter, and the real-time position of the slag tank trolley is calculated. By controlling the motor, the slag tank trolley is driven to the target position, so that the slag tank on the slag tank trolley can take over the converter in real time. Slag

   When the difference between the total weight of dumped slag G _t and the target amount of dumped slag G ₀ is the preset deviation amount, stop the converter tilting and start the converter rollback.
2. The method for automatic slag dumping of a converter slagging process based on image recognition according to claim 1, wherein the real-time image of the hot slag flow during the converter slag dumping process is collected and stored, and the real-time image includes : One of visible light image, infrared image and far-infrared image.
3. The method of automatic image recognition based deslagging the slag leaving the converter process according to claim 1, characterized in that the volume V _i and the time-based deslagging each time the image recognition calculation is the tracking region slag In the total weight G _t of slag dumping at t, the following content is used to calculate the volume of slag V _{i at each moment in the tracking area and the total weight G t of} slag dumping when the slag dumping time is t based on image recognition:

   Divide the tracking area into i×j grids, scan the tracking area line by line with a scan width Δh, and the intersection of the feature boundary and the grid in each scan width forms a micro area;

   Calculate the slag thickness based on the following model:

   

   Where δ _j represents the thickness of the slag within a scan width at any time;

   

   Represents the average width of the slag within a sweep width at any time; α represents the coefficient related to the slag viscosity, the greater the slag viscosity, the greater the value of α, and its value range is 0＜α≤1; λ is the coefficient and its value range 0＜λ≤1; n is a coefficient, and its value range is 0＜n≤2;

   Then in the micro area with the sweep width of △h, the volume of the slag is

   

   Sum the volume of each micro area to obtain the volume V _i of the slag in the tracking area at the current moment;

   When it is detected that the volume of slag in the tracking area _Vi > the set detection threshold V _min , the time at this time is marked as 0, and when the slag dumping time reaches t, the total weight G _{t of the} slag dumping is based on the following The formula of the model is obtained:

   

   Where ρ _slag represents the density of the slag, and g represents the acceleration due to gravity.
4. The method for automatic slag dumping of converter slag retention process based on image recognition according to claim 1, characterized in that:

   When the volume of slag in the tracking area _Vi ≤ the set detection threshold V _min , control the tilting angular velocity of the converter to the set first angular velocity ω ₁ , at this time ω _i =ω ₁ ;

   When the volume of slag in the tracking area _Vi > the set detection threshold V _min , control the tilting angular velocity of the converter to the set second angular velocity ω ₂ , and ω ₂ <ω ₁ , at this time ω _i = ω ₂ ;

   When the volume of slag in the tracking area _Vi > the set detection threshold V _min , and the total weight of dumped slag G _t satisfies

   

   And when |G _t -G ₀ |>G _min , control the real-time tilting angular velocity of the converter

   

   Where t represents the converter slag dumping time; G _min represents the allowable deviation of the slag weight, which is the set value; k ₁ represents the change rate of the slag weight; b represents the adjustment coefficient, and its value range is 0＜b≤1; The total weight of slag G _t satisfies

   

   And when |G _t -G ₀ |>G _min , control the real-time tilting angular velocity of the converter ω _i =0.
5. The method for automatic slag dumping of the converter slag retention process based on image recognition according to claim 4, characterized in that the real-time position x _{i of the} slag tank trolley is controlled according to the following model:

   x _i = x ₀ + βθ _i

   Among them, x ₀ represents the initial position of the slag tank trolley; β represents the adjustment coefficient, which ranges from 0.01 to 0.1, and the unit parameter is meters/degree (m/°); θ _i = θ ₀ +ω _i t, Where θ _i represents the real-time tilting angle of the _{converter, θ 0} represents the initial tilting angle of the converter, t represents the slag _{dumping time of the converter, and ω i} is the real-time tilting angular velocity of the converter.
6. An automatic slag dumping system for converter slag retention process based on image recognition, which is characterized in that it includes:

   Converter

   The converter tilting driving device is connected with the converter to drive the converter to tilt and perform the slag dumping action;

   Converter tilt angle measuring device, used to measure the tilt angle of the converter;

   The slag tank is set on the slag tank trolley;

   The slag tank trolley position detection device is used to detect the real-time position of the slag tank trolley;

   Image acquisition device for acquiring real-time images of the hot slag flow during the slag dumping process of the converter;

   An image processing module for processing the real-time image;

   The storage module is used to store the real-time image and/or the real-time image processed by the image processing module;

   The image recognition and calculation module is used to recognize the real-time image to determine the characteristic boundary of the hot slag flow, and calculate the volume of the slag in the selected tracking area in real time;

   The control module is respectively connected with the converter tilting drive device, the converter tilt angle measuring device, the slag tank trolley, the slag tank trolley position detection device, the image recognition and calculation module, and the control module is set to: calculate based on image recognition when deslagging is t region of each track within a selected time slag volume V _i and the total weight G deslagging time _t; slag when the trace in the region of the volume V _i> set detection threshold V _min, Based on the weight change rate dG _t /dt, the real-time tilting angular velocity ω _{i of the} converter is controlled, and the real-time tilting angle of the converter is obtained according to the real-time tilting angular velocity, and the real-time position of the slag tank trolley is calculated based on the real-time tilting angle of the converter , So that the slag tank can accept the slag poured from the converter in real time; when _{the difference between the total weight of the slag dumped G t} and the target slag _{dumping amount G 0} is the preset deviation amount, stop the converter tilting and start the converter rollback.
7. The system for automatic slag dumping of the converter slag retention process based on image recognition according to claim 6, characterized in that it further comprises a production process computer system connected to the control module, and the production process computer system sends the converter to the control module. Smelting data.
8. The system for automatic slag dumping of converter slag retention process based on image recognition according to claim 6, wherein the image acquisition device includes one of a visible light camera, an infrared camera, and a far infrared camera.
9. The system for automatic slag dumping of a converter slag retention process based on image recognition according to claim 6, wherein the converter inclination angle measuring device comprises an encoder or an inclinometer.
10. The system for automatic slag dumping of converter slag retention process based on image recognition according to claim 6, wherein the slag tank trolley position detection device includes a non-contact distance meter.

Images