A dual-tap hole blast furnace tap hole area dust removal power saving control system and method

Abstract

The present application relates to a kind of double iron notch blast furnace casthouse dust removal power saving control system and method, belong to welding equipment technical field.The technical scheme is: when No.1 iron notch (3) and / or No.2 iron notch (4) is in the non-iron stage, control unit controls the opening of valve one (1) and / or valve two (2) reduces, when No.1 iron notch (1) and / or No.2 iron notch (2) is in the iron stage, control unit controls the opening of valve one (1) and / or valve two (2) increases;Fan (5) is connected in the pipeline between valve one (1) and valve two (2), the power of fan (5) is concentrated in the valve one (1) or valve two (2) where more smoke dust, while control unit controls different speed gears of fan (5) by frequency converter.The present application realizes the optimal matching of fan dust removal power and blast furnace gas volume, eliminates the problem of energy waste caused by excessive fan speed and the problem of smoke dust overflow caused by low fan speed.

Description

TECHNICAL FIELD

[0001] The present application relates to a double-tap hole blast furnace tap hole yard dust removal power saving control system and method, belonging to the technical field of welding equipment. BACKGROUND

[0002] In the process of producing molten iron, a large amount of smoke dust will be generated with the opening and closing of the tap hole and the flow of molten iron, so a dust removal power saving control system needs to be set. The number of tap holes of a blast furnace varies according to the specifications of the blast furnace, and a common blast furnace is provided with two tap holes. For a blast furnace with two tap holes, according to the requirements of the ironmaking process, single tap hole tapping is generally allowed, and double tap hole tapping is allowed when the production load is high.

[0003] So far, the dust removal system of the blast furnace tap hole is mostly manually controlled, and the operation process is roughly as follows: when the amount of smoke dust of the tap hole increases or decreases obviously, the operator controls the frequency of the fan through the on-site image, and when the state of the tap hole changes from tapping to non-tapping or from non-tapping to starting tapping, the operator controls the frequency of the fan and the opening of the valve through the on-site image. However, since the operator needs to control multiple programs at the same time, it is easy to cause smoke leakage and power waste due to untimely control signals. Therefore, how to design a dust removal power saving control system so that the dust removal power saving control system can automatically detect the amount of smoke dust and timely adjust to different dust removal operation modes is a technical problem that needs to be solved by those skilled in the art. SUMMARY

[0004] The purpose of the present application is to provide a double-tap hole blast furnace tap hole yard dust removal power saving control system and method, which can timely and accurately adjust the fan speed and valve opening according to the amount of smoke dust, meet the dust removal needs under various working conditions such as tap hole tapping and non-tapping working conditions, and solve the above problems existing in the background art.

[0005] The technical scheme of the present application is:

[0006] The application discloses a kind of dual-tap hole blast furnace tap hole yard dust removal power-saving control systems, include visual unit, control unit, fan and valve unit;Wherein: visual unit is responsible for monitoring the smoke dust amount of No. 1 tap hole and No. 2 tap hole, and different state signals are output to control unit according to the smoke dust amount, and control unit receives state signal, and valve unit and fan unit are controlled to respond respectively;Valve unit is divided into two parts, and valve one and valve two are arranged in the dust removal pipeline of No. 1 tap hole and No. 2 tap hole respectively, when No. 1 tap hole and / or No. 2 tap hole is in non-tapping stage, control unit controls the opening of valve one and / or valve two to reduce, when No. 1 tap hole and / or No. 2 tap hole is in tapping stage, control unit controls the opening of valve one and / or valve two to increase;Fan unit includes a fan, which is connected in series in the pipeline between valve one and valve two, and control unit controls the power of fan to concentrate in valve one or valve two with more smoke, to improve fan utilization rate, and control unit controls different speed gears of fan through frequency converter to match different smoke amounts.

[0007] Further, the visual unit includes camera one and camera two, which are respectively installed at No. 1 tap hole and No. 2 tap hole, and are respectively used for monitoring the smoke dust amount of the first tap hole and the second tap hole.

[0008] Further, the visual unit includes a computer, which is used for receiving image signals of camera one and camera two, inputting the signals to a python program for image processing, and transmitting state information obtained after processing to the control unit.

[0009] Further, the control unit includes a PLC controller and an HMI human-computer interaction interface, and has two operation modes, namely an automatic mode and a manual mode.

[0010] Further, the fan is a dust removal fan, which is a publicly known technology and can be purchased on the market.

[0011] In normal production process, the system will be in automatic mode, in which mode, the PLC controller directly receives state signals transmitted by the computer, and adjusts the fan unit and the valve unit to a preset state according to different state signals.

[0012] When abnormal condition occurs or in the case of need, it can be switched to manual mode, in which mode, the PLC controller no longer receives computer signals, and only accepts human control of the HMI human-computer interaction interface. Controllable contents include fan frequency and the opening of valve one and valve two in the dust removal pipeline of No. 1 tap hole and No. 2 tap hole.

[0013] Further, the fan is installed behind the intersection of the dust removal pipeline of the No. 1 and No. 2 tapping holes, directly acts on the dust removal pipeline of the tapping holes, generates airflow, and sucks and transports the smoke dust in the dust removal pipeline of the tapping holes to the dust removal equipment.

[0014] The fan has four basic gears in the automatic mode, namely a low speed gear, a medium speed gear, a high speed gear and an ultra-high speed gear.

[0015] Further, the valve one and the valve two are main control valves, and the valve one and the valve two are connected with eight subordinate valves at corresponding positions of the No. 1 and No. 2 tapping holes respectively.

[0016] A dust removal and power saving control method for a double-tapping hole blast furnace tapping field, which is performed by using the above-mentioned dust removal and power saving control system for a double-tapping hole blast furnace tapping field.

[0017] (1) The states of the No. 1 and No. 2 tapping holes are set as a tapping stage and a non-tapping stage:

[0018] ① The tapping stage includes two periods: a tapping smoke dust period in which the state of a large amount of smoke dust during tapping is divided, and the fan is in a high speed gear or an ultra-high speed gear; a tapping stable period in which the state of a small amount of smoke dust during tapping is divided, and the fan is in a medium speed gear;

[0019] ② The time period from the end of tapping to the beginning of tapping in the next cycle is divided into a non-tapping stage, and the fan is in a low speed gear;

[0020] (2) The state of the No. 1 and No. 2 tapping holes is detected by an automatic tapping detection function, and the specific steps are as follows:

[0021] The visual unit is responsible for monitoring the amount of smoke dust of the No. 1 and No. 2 tapping holes, and transmitting information to the control unit, and the control unit processes the signal of the visual unit to determine the state of the No. 1 and No. 2 tapping holes;

[0022] The opening degree of the valve one and the valve two and the gear of the fan are controlled by the control unit, and the specific steps are as follows:

[0023] ① When the visual unit detects that the No. 1 tapping hole is in the tapping stage and the No. 2 tapping hole is in the non-tapping stage, the control unit adjusts the opening degree of the valve one to a high conduction state, the opening degree of the valve two to a low conduction state, and the power of the fan is concentrated on the valve one where the No. 1 tapping hole is located, and the speed of the fan is adjusted to the corresponding gear according to the amount of smoke dust at the No. 1 tapping hole;

[0024] ② When the vision unit detects that the No. 1 iron outlet is in the non-iron outlet stage and the No. 2 iron outlet is in the iron outlet stage, the opening of the unit regulating valve one is adjusted to the low conduction state and the opening of valve two is adjusted to the high conduction state. The power of the blower is concentrated at valve two where the No. 2 iron outlet is located. At the same time, the speed of the blower is adjusted to the corresponding gear according to the amount of smoke and dust at the No. 2 iron outlet.

[0025] ③ When the vision unit detects that both No. 1 and No. 2 iron outlets are in the iron outlet stage, the opening of valves 1 and 2 is adjusted to the high conduction state, and the speed of the blower is controlled to be kept at the corresponding gear until No. 1 and No. 2 iron outlets no longer meet the conditions for the iron outlet stage.

[0026] ④ When the vision unit detects that both No. 1 and No. 2 iron outlets are in the non-iron outlet stage, the opening of valves 1 and 2 is adjusted to the low conduction state, and the fan speed is controlled to be maintained at the corresponding gear.

[0027] Furthermore, the automatic iron detection function is implemented through deep learning recognition methods or marker recognition methods:

[0028] The deep learning recognition method is as follows: by collecting iron tapping images from three stages of iron tapping, a dataset was constructed and a model was trained. This model can accurately identify the state of iron tapping point 1 and iron tapping point 2.

[0029] The marker identification method is as follows: install a marker plate behind the smoke and dust at the No. 1 and No. 2 iron outlets, and use OpenCV to identify the green area of ​​the marker plate to determine the amount of smoke and dust, so as to accurately identify the status of the No. 1 and No. 2 iron outlets.

[0030] The deep learning recognition method and the marker recognition method can be run individually or together to complete the task of identifying the amount of smoke and dust, with the recognition result of the larger amount of smoke and dust being used as the standard.

[0031] Furthermore, based on the deep learning recognition method and the marker recognition method to identify the status of the No. 1 and No. 2 iron taps, a separate neural network model is added to identify worker operations; if there are workers operating at the No. 1 and No. 2 iron taps, the current fan speed is increased by a preset value; after the workers finish their work and leave the site, the fan speed will return to the base speed.

[0032] Further, when the visual unit determines that the current state is the non-tapping stage and there is no worker operation, first, the position of the identification plate is profile-recognized and recalibrated using OpenCV, and the color and brightness of the identification plate are recognized; if below a preset value, the operator is prompted to wipe through the HMI indicator light, and it is suggested that the worker switches to only be judged by the neural network method, or is temporarily adjusted to the manual mode.

[0033] The positive effect of the present application: the fan speed and valve opening degree can be adjusted in time and accurately according to the size of smoke dust, the dust removal demand under various working conditions such as tapping working condition and non-tapping working condition is met, automatic regulation and control of double tapping hole dust removal is realized, the optimal matching of fan dust removal power and tapping field smoke amount is realized while reducing the labor intensity of manual operation, the problems of energy waste caused by too large fan speed and smoke dust overflow caused by too low fan speed are eliminated. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 It is a structural schematic diagram of the embodiment of the present application;

[0035] Figure 2 It is an HMI man-machine interface of the embodiment of the present application;

[0036] Figure 3 It is a valve opening degree and motor frequency change curve of the embodiment of the present application;

[0037] Figure 4 It is a flow chart of the embodiment of the present application;

[0038] Figure 5 It is a schematic diagram of the automatic mode principle of the embodiment of the present application;

[0039] Figure 6 It is a schematic diagram of the manual mode principle of the embodiment of the present application;

[0040] In the figure: valve one 1, valve two 2, No. 1 tapping hole 3, No. 2 tapping hole 4, fan 5, camera one 6, camera two 7, computer 8, PLC controller 9, HMI man-machine interface 10, identification plate 11, groove 12. DETAILED DESCRIPTION

[0041] The present application will be further described below in combination with the drawings and embodiments:

[0042] A dust removal and energy-saving control system for a blast furnace tapping area with dual tapholes includes a vision unit, a control unit, a fan, and a valve unit. The vision unit monitors the dust levels at tapholes 1 and 2 and outputs different status signals to the control unit based on the dust levels. The control unit receives these status signals and controls the valve unit and fan unit accordingly. The valve unit consists of two parts: valve 1 (at taphole 3) and valve 2 (at taphole 4) installed in the dust removal pipes at tapholes 1 and 2. When taphole 3 and / or taphole 4 are in an uncontrolled state... During the tapping stage, the control unit controls the opening of valve 1 and / or valve 2 to decrease. When tapping port 1 and / or tapping port 2 are in the tapping stage, the control unit controls the opening of valve 1 and / or valve 2 to increase. The fan unit includes a fan 5, which is connected in series in the pipe between valve 1 and valve 2. The control unit controls the power of fan 5 to concentrate at valve 1 or valve 2 where there is more smoke and dust, thereby improving the fan utilization rate. At the same time, the control unit controls different speed levels of fan 5 through a frequency converter to match different amounts of smoke and dust.

[0043] The vision unit includes a first camera 6 and a second camera 7, which are installed at the first iron outlet 3 and the second iron outlet 4, respectively, and are used to monitor the amount of smoke and dust at the first iron outlet 3 and the second iron outlet 4.

[0044] The vision unit includes a computer 8, which needs to be equipped with a Python environment and have the OpenCV library and other libraries required for deep learning installed. The computer 8 is used to receive image signals from camera 6 and camera 7, input the signals to the Python program for image processing, and transmit the processed status information to the control unit.

[0045] The control unit includes a PLC controller 9 and an HMI human-machine interface 10, and has two operating modes: automatic mode and manual mode.

[0046] During normal production, the system will be in automatic mode. In this mode, PLC controller 9 will directly receive status signals from the computer and adjust the fan unit and valve unit to preset states according to different status signals. In case of abnormal conditions or when necessary, it can be switched to manual mode. In this mode, PLC controller 9 will no longer receive computer signals and will only accept manual control through the HMI (Human-Machine Interface).

[0047] The fan 5 is installed behind the intersection of the dust removal pipes of No. 1 iron outlet 3 and No. 2 iron outlet 4, and directly acts on the dust removal pipes of the iron outlets to generate airflow, which sucks up the smoke and dust in the dust removal pipes of the iron outlets and transports them to the dust removal equipment.

[0048] The fan has four basic speed settings in automatic mode: low speed, medium speed, high speed, and ultra-high speed. When needed on site, a fixed preset value can be added to the basic speed settings.

[0049] Both valve 1 and valve 2 are main control valves. Valve 1 and valve 2 are respectively connected to eight downstream valves at corresponding positions at taphole 3 and taphole 4. The distribution of the eight downstream valves is as follows: one valve each at the small pit, slag ditch, residue ditch, and taphole; one valve each on the north and south sides of the slag ditch; and one valve each on the north and south sides of the ladle position. The above eight downstream valves are installed in the corresponding positions of the dust removal pipeline at the taphole.

[0050] In this embodiment, the control unit only controls valve one and valve two, and the lower-level valves are manually adjusted by on-site personnel according to production conditions. Valve one and valve two, along with the eight matched lower-level valves, constitute a valve unit. In automatic mode, the valve unit has two conduction states: a low opening state with a preset value of 20% opening and a high opening state with a preset value of 80% opening.

[0051] A method for dust removal and energy saving control in the tapping area of ​​a blast furnace with dual tapping outlets, employing the aforementioned dust removal and energy saving control system for the tapping area of ​​a blast furnace with dual tapping outlets:

[0052] (1) Set the states of tapping port 3 and tapping port 4 to tapping stage and non-tapping stage respectively:

[0053] ① The tapping stage includes two periods: the tapping dust period is defined as the state with a large amount of smoke and dust when tapping iron, during which the fan 5 is at a high speed or ultra-high speed to ensure the dust removal effect; the tapping stable period is defined as the state with a small amount of smoke and dust when tapping iron, during which the fan 5 is at a medium speed to save electricity.

[0054] ② The time period from the end of tapping iron to the start of the next cycle is divided into the non-tapping iron stage. During this time, the blower 5 is in a low speed position to save electricity.

[0055] (2) The status of tapping port 3 and tapping port 4 is detected by the automatic tapping detection function. The specific steps are as follows:

[0056] The vision unit is responsible for monitoring the amount of smoke and dust at tap 3 and tap 4 and transmitting the information to the control unit. The control unit processes the signal from the vision unit and determines the status of tap 3 and tap 4.

[0057] (3) The opening degree of valve one and valve two and the speed of the blower are controlled by the control unit. The specific steps are as follows:

[0058] ① When the vision unit detects that the No. 1 iron outlet 3 is in the iron-producing stage and the No. 2 iron outlet 4 is in the non-iron-producing stage, the control unit adjusts the opening of valve 1 to a high conduction state and the opening of valve 2 to a low conduction state. The power of the fan 5 is concentrated at valve 1 where the No. 1 iron outlet 3 is located. At the same time, the fan speed is adjusted to the corresponding gear according to the amount of smoke and dust at the No. 1 iron outlet (3).

[0059] ② When the vision unit detects that the No. 1 iron outlet 3 is in the non-iron outlet stage and the No. 2 iron outlet 4 is in the iron outlet stage, the opening of the unit regulating valve 1 is adjusted to the low conduction state, the opening of valve 2 is adjusted to the high conduction state, the power of the fan 5 is concentrated at valve 2 where the No. 2 iron outlet 4 is located, and the fan speed is adjusted to the corresponding gear according to the amount of smoke and dust at the No. 2 iron outlet 4.

[0060] ③ When the vision unit detects that both No. 1 iron outlet 3 and No. 2 iron outlet 4 are in the iron-exit stage, the opening of valve 1 and valve 2 is adjusted to the high conduction state, and the speed of the fan 5 is controlled to be kept at the corresponding gear until No. 1 iron outlet 3 and No. 2 iron outlet 4 do not meet the conditions for the iron-exit stage.

[0061] ④ When the vision unit detects that both No. 1 iron outlet 3 and No. 2 iron outlet 4 are in the non-iron outlet stage, the opening of valve 1 and valve 2 is adjusted to the low conduction state, and the speed of the fan 5 is controlled to be kept at the corresponding gear until the conditions that No. 1 iron outlet 3 and No. 2 iron outlet 4 do not meet the non-iron outlet stage are not met are not met.

[0062] The table below shows the operating conditions of the fan, valve one, and valve two;

[0063]

[0064] The automatic iron detection function is implemented using deep learning recognition methods and marker recognition methods:

[0065] The deep learning recognition method is as follows: by collecting iron tapping images from three stages of iron tapping, a dataset is constructed and a model is trained. This model can accurately identify the state of iron tapping point 1 and iron tapping point 2.

[0066] The deep learning recognition method described uses TensorFlow to build the model framework. The model includes an input layer, convolutional layers, pooling layers, fully connected layers, an output layer, and activation functions. The input layer receives raw image information from the camera in the visual unit. The input image first passes through three convolutional layers, each containing multiple convolutional kernels that slide across the image to detect different features. The convolution operation generates a series of feature maps, each corresponding to a convolutional kernel, representing the location and intensity of the feature in the image. A ReLU activation function is used after each convolutional layer to introduce non-linearity, enabling the model to learn very complex features. Each convolutional layer is followed by a max-pooling layer to reduce the spatial dimensionality of the feature maps, reducing computational cost. After the convolutional and pooling layers, the model includes a fully connected layer that flattens the feature maps into a one-dimensional vector. This fully connected layer contains 128 neurons and is used to learn higher-level feature representations. The output layer contains three neurons, corresponding to three categories: the iron-extraction dust period, the iron-extraction plateau period, and the no-iron-extraction period. The output layer uses the softmax activation function to calculate the probability distribution for each category. The model then classifies the input image into the category with the highest probability.

[0067] The deep learning-based recognition method includes a deep learning model compilation process. This method uses the Adam optimizer to tune model parameters to minimize loss, the categorical_crossentropy loss function to measure the difference between the model's predictions and the actual labels, and accuracy to monitor the model's performance during training. These configurations determine the model training method and performance evaluation criteria, providing guidance for model training.

[0068] The deep learning recognition method includes an image type recognition process. The image transmitted by the visual unit is input into an image classification function, where a model is used to classify the image, and the output image belongs to one of three categories: the iron tapping dust period, the iron tapping stable period, or no iron tapping.

[0069] The marker identification method is as follows: green identification plates are installed behind the smoke and dust at the No. 1 and No. 2 iron tapping ports. The amount of smoke and dust is determined by the green area of ​​the identification plates using OpenCV, so as to accurately identify the status of the No. 1 and No. 2 iron tapping ports.

[0070] The marker recognition method includes a brightness adjustment function. This function converts the input image into a grayscale image to calculate the brightness. The processed image undergoes average brightness detection, and the difference between the average brightness and the desired brightness level is calculated. Then, the `cv2.convertScaleAbs` function is used to adjust the brightness while maintaining the image contrast. Finally, the adjusted image is returned to the calling function.

[0071] The marker recognition method includes a green area calculation function. This function receives the processing output of the brightness adjustment function, defines a rectangular area of ​​the marker, and calculates its area. Next, the image within the area is converted to the HSV color space, and the HSV range of the green color is defined for color recognition. Then, a binary mask is used to turn pixels matching the range white and other pixels black. The coordinates of the white pixels are added to a tuple, and the area of ​​the green region is calculated by calculating the length of the tuple.

[0072] The marker recognition method includes a category determination function. This function defines the ratio of the number of green pixels (n) to the total number of pixels (m) in the region as the green occupancy ratio (α). The calculation formula is:

[0073] α=

[0074] The green occupancy ratio is compared with preset values. If the green occupancy ratio is greater than the preset value 1, the current state is determined to be no iron tapping. If the green occupancy ratio is greater than the preset value 2 and less than the preset value 1, the current state is determined to be the iron tapping stable period. If the green occupancy ratio is less than the preset value 2, the current state is determined to be the iron tapping flue dust period.

[0075] The two identification methods described above can be used individually or together to complete the identification task, with the result of the larger smoke and dust volume being used as the standard. The selection of the identification method is set in the HMI of the control unit and is controlled by the operator.

[0076] Regardless of whether the system is in automatic or manual mode, the recognition results of both recognition methods will be transmitted to the PLC of the control unit.

[0077] When the system is in automatic mode, regardless of which of the two identification methods is selected to run, or whether both methods are run, the identification result will be transmitted to the PLC of the control unit.

[0078] When the system selects to run the deep learning recognition method, the PLC processes the recognition results of the deep learning recognition method, but does not process the recognition results of the marker recognition method.

[0079] When the system selects to run the marker recognition method, the PLC processes the recognition results of the marker recognition method, while the recognition results of the deep learning recognition method are not processed.

[0080] Here, the depth recognition method is abbreviated as A, and its dust volume determination result is a; the marker recognition method is abbreviated as B, and its dust volume determination result is b. According to the above dust volume classification rules, the dust volume during the pre-tapping stage < the dust volume during the stable tapping stage < the dust volume during the tapping stage. When the system selects both methods to run, the PLC processing logic is as follows:

[0081] When a = b, process the recognition result of A; when a > b, process the recognition result of A; when a < b, process the recognition result of B.

[0082] Based on the recognition of the states of the No. 1 and No. 2 tapholes by the deep learning recognition method and the marker recognition method, a separate neural network model is added to recognize the worker operations. If there are worker operations in the current environment, a signal to increase the rotational speed is sent to the control unit, and the rotational speed of the fan is increased by a preset value based on the current state. When the workers finish their operations and leave the site, the vision unit sends a signal to stop increasing the rotational speed, and the fan rotational speed will return to the basic rotational speed.

[0083] The marker recognition method has a function of detecting the state of the identification board and a function of calibrating the position of the identification board.

[0084] Function of detecting the state of the identification board: When the vision unit determines that the current state is the non-tapping stage and there is no worker operation, recognize the color and brightness of the identification board; if it is lower than a preset value, prompt the operator to wipe it through the HMI indicator light, and suggest that the staff switch to only judge by the neural network method or temporarily adjust to the manual mode.

[0085] Function of calibrating the position of the identification board: When the vision unit determines that the current state is the non-tapping stage and there is no worker operation, OpenCV performs contour recognition and recalibration on the position of the identification board. The realization of this function first requires reading the image containing the identification board and converting the image to the HSV color space, and defining an HSV color range for green to create a mask to extract the area of the identification board in the image. Then use the contour detection method to find the contours in the extracted green area and obtain their coordinate information for the recalibration operation.

[0086] Refer to Attachment Figures 1-6 , this embodiment provides a dust removal and power saving control system for the tapping yard of a blast furnace with two tapholes.

[0087] Refer to Attachment Figure 1 , the two tapholes of the blast furnace equipment with two tapholes are the No. 1 taphole 3 and the No. 2 taphole 4 respectively.

[0088] There are two cameras, both of which use the type I pinhole cameras in Hikvision, and are respectively installed at the No. 1 taphole 3 and the No. 2 taphole 4 to monitor the states of the No. 1 taphole 3 and the No. 2 taphole 4. This system includes a computer equipped with an i7, which needs to install a python environment and an HMI human-machine interface, to receive the image signals of the cameras, input the signals to the python program for image processing, and transmit the processed state information to the CPU.

[0089] It also includes a Siemens S7-1511 CPU and an HMI (Human Machine Interface), the HMI screen of which is as follows: Figure 2 As shown, the display content includes the current tapping status of tapping outlet 3 and tapping outlet 4, the selection and display of the tapping status identification method, whether there are workers operating on site, the current valve unit opening control and display, the fan unit frequency control and display, and the selection and display of the two operating modes.

[0090] The dust removal and energy-saving control system provided in this embodiment has two operating modes: automatic mode and manual mode. During normal production, the system will operate in automatic mode. In this mode, the PLC controller 9 will directly receive status signals from the computer and adjust the fan unit and valve unit to preset states according to different status signals, displaying the current values ​​on the HMI screen. In case of abnormal conditions or when necessary, it can be switched to manual mode. In this mode, the PLC controller 9 will no longer receive computer signals and will only accept manual control via the HMI human-machine interface. Controllable parameters include fan frequency and valve opening of each dust removal pipe at the iron outlet.

[0091] The dust removal and energy-saving control system provided in this embodiment includes one dust removal fan (fan 5) and two valves (valve 1 and valve 2). Using a dual-fan system requires more installation space, which may limit its applicability, especially in confined spaces in industrial environments. Secondly, since a dual-fan system involves two independently operating fans, a more complex control system is needed to ensure their coordinated operation. Furthermore, if the two fans do not work in coordination or are improperly controlled, incomplete dust removal may occur, thus affecting environmental pollution control compliance.

[0092] In comparison, the single-fan, dual-valve system used in this system reduces energy consumption and operating costs. Furthermore, with only one dust collector fan, the system requires less installation space, making it suitable for industrial environments with limited space. The system is also simpler to control, reducing complexity and maintenance difficulty. Moreover, the single-fan, dual-valve system allows for more precise control to meet environmental regulations and standards. In conclusion, the single-fan, dual-valve system offers a more economical and effective solution in specific situations.

[0093] The fan unit is installed behind the intersection of the dust removal pipes of the two iron outlets. It directly acts on the main dust removal pipe to generate airflow. The air mixed with dust in the No. 1 and No. 2 iron outlets will enter the main dust removal pipe from the branch pipes with valves on their respective iron outlets, and will be transported to the dust removal equipment by the fan. After meeting the environmental protection requirements, it will be discharged into the atmosphere through the chimney.

[0094] The fan unit has four basic speed settings in automatic mode: low speed, medium speed, high speed, and ultra-high speed. A fixed preset value can be added to the basic settings when needed on-site.

[0095] The valve unit consists of two valves (Valve 1 and Valve 2), which are installed on the dust collection pipes of taphole 1 and taphole 2, respectively. Both Valve 1 and Valve 2 are master control valves, each associated with eight subordinate valves: one valve each for the small pit, slag ditch, residue ditch, and taphole; one valve on each side of the slag ditch; and one valve on each side of the ladle position. The control program only controls Valve 1 and Valve 2; the subordinate valves are manually adjusted by on-site personnel according to production conditions. In automatic mode, the valve unit has two on / off states: a low opening state with a preset value of 20% and a high opening state with a preset value of 80%.

[0096] The computer records the operating status of the fan unit and valve unit every 2 seconds and displays it on the screen in the form of a line graph, as shown below. Figure 3 As shown.

[0097] like Figure 4 As shown in the figure, this embodiment provides a dust removal and energy-saving control method for the tapping area of ​​a blast furnace with dual tapping outlets, including automatic mode and manual mode. The specific process is as follows:

[0098] During normal production, the system will be in automatic mode. In this mode, PLC controller 9 will directly receive status signals transmitted from the computer and adjust the fan unit and valve unit to preset states according to different status signals. Camera 1 and Camera 2 in the vision unit will detect the iron tapping ports 1 and 2 that are currently tapping iron, respectively, and transmit the information to the control unit. The control unit will process the signals from the vision unit, determine the status of the two iron tapping ports, and control the corresponding units to make adjustments.

[0099] This embodiment provides two methods for visual units to identify the state of iron tapping.

[0100] A deep learning-based recognition method was developed by collecting images of iron tapping at various stages of the tapping process, constructing a dataset, and training a model that can accurately identify the state of the tapping. The deep learning recognition method uses TensorFlow to build the model framework. The model includes an input layer, convolutional layers, pooling layers, fully connected layers, an output layer, and activation functions. The model classifies the input image into the category with the highest probability.

[0101] During the recognition process, the image transmitted by the visual unit will be input into the image classification function to be classified by the model, and the output image will belong to one of the three categories: iron tapping dust period, iron tapping stable period, and no iron tapping.

[0102] The marker recognition method involves installing markers behind smoke and dust, and using OpenCV to identify the green area of ​​the markers to determine the amount of smoke and dust. The method converts the input image to grayscale to calculate brightness. It then performs average brightness detection on the processed image and calculates the difference between the average brightness and the desired brightness level. Next, the `cv2.convertScaleAbs` function is used to adjust the brightness while maintaining image contrast. A designated area for the markers is then defined, and its area is calculated. The image within this area is converted to the HSV color space, and the HSV range for green is defined for color recognition. A binary mask is then used to turn pixels within the range white and other pixels black. The coordinates of the white pixels are included in a tuple, and the area of ​​the green region is calculated by calculating the length of the tuple. This function defines the ratio of the number of green pixels to the total number of pixels in the region as the green occupancy ratio. The green occupancy ratio is compared with preset values. If the green occupancy ratio is 1, the current state is determined to be no iron being tapped. If the green occupancy ratio is greater than the preset value 2 and less than the preset value 1, the current state is determined to be the iron tapping stable period. If the green occupancy ratio is less than the preset value 2, the current state is determined to be the iron tapping flue gas period.

[0103] The two identification methods can be run individually or together to complete the identification task, with the identification result of the larger amount of smoke and dust being used as the standard.

[0104] The dust removal and energy-saving control method provided in this embodiment includes a worker operation detection function. Based on the two identification methods mentioned above for recognizing the status of taphole 1 and taphole 2, a separate neural network model is added for worker operation identification. If worker operation is present, the current fan speed is increased by a preset value. After the worker completes their work and leaves the site, the fan speed will return to its base speed.

[0105] The dust removal and energy-saving control method provided in this embodiment includes a label plate status detection function. When the vision unit determines that the current state is no iron being tapped and no workers are operating, it identifies the color and brightness of the label plate. If the brightness is below a preset value, the operator is prompted to wipe the label plate via an HMI indicator light, and it is suggested that the operator switch to a neural network-based method for judgment, or temporarily switch to manual mode.

[0106] The dust removal and energy-saving control method provided in this embodiment includes a marker plate position calibration function. When the vision unit determines that the current state is no iron being tapped and no workers are operating, OpenCV is used to perform contour recognition on the position of the marker plate and recalibrate it.

[0107] In case of abnormal situations or when necessary, you can switch to manual mode, such as... Figure 6 As shown, in this mode, the PLC controller will no longer receive computer signals, but will only accept manual control via the HMI (Human-Machine Interface). Controllable parameters include fan frequency and the opening degrees of valves one and two.

[0108] The dust removal and energy-saving control methods provided in this embodiment include single iron outlet iron discharge strategy, dual iron outlet iron discharge strategy, and iron discharge strategy without iron outlet.

[0109] The single-outlet tapping strategy is as follows: When the vision unit detects that outlet 1 is in the tapping stage and outlet 2 is not in the tapping stage, the control unit adjusts valve 1 to a high-conductivity state and valve 2 to a low-conductivity state, while simultaneously adjusting the fan speed to the corresponding gear according to the tapping stage of outlet 1. When the vision unit detects that outlet 2 is in the tapping stage and outlet 1 is not in the tapping stage, the control unit adjusts valve 2 to a high-conductivity state and valve 1 to a low-conductivity state, while simultaneously adjusting the fan speed to the corresponding gear according to the tapping stage of outlet 2.

[0110] The dual-outlet iron discharge strategy is as follows: when the vision unit detects that both outlet 1 and outlet 2 are in the iron discharge stage, it will adjust valve 1 and valve 2 to be in a high conduction state and control the fan speed to be kept at an ultra-high speed until outlet 1 and outlet 2 no longer meet the iron discharge conditions.

[0111] The iron discharge strategy without iron outlets: When the vision unit detects that both iron outlet No. 1 and iron outlet No. 2 are in the iron-discharging stage, it will adjust valves 1 and 2 to a low conduction state and control the fan speed to remain at a low speed until neither iron outlet No. 1 nor iron outlet No. 2 meets the conditions for not discharging iron.

[0112] In summary, the dust removal and energy-saving control system and method for the tapping area of ​​a blast furnace with two tapping outlets provided in this embodiment, by setting up two valves, one fan, two cameras, and one computer, meets the automatic dust removal requirements of the blast furnace equipment under the conditions of tapping iron at both tapping outlets, tapping iron at one tapping outlet, and tapping iron at no tapping outlet. This system can switch between automatic and manual modes at any time, and in automatic mode, three recognition methods are available for selection. It adjusts the fan speed and valve opening accurately and promptly according to the amount of smoke and dust, meeting the dust removal requirements under various conditions such as tapping iron at the tapping outlet and no tapping iron, achieving automatic control of dust removal at both tapping outlets, reducing the need for manual operation while avoiding excessive energy consumption.

[0113] The above examples illustrate the principles and implementation methods of this embodiment.

[0114] This is merely intended to aid in understanding the method and core ideas of this embodiment. It should be noted that those skilled in the art can make various improvements and modifications to this embodiment without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this embodiment.

Claims

1. A dual-tap hole blast furnace casthouse dust removal and power saving control system, characterized in that: The application relates to a double-outlet-tap blast furnace tap hole dust removal and power saving control system, which comprises a visual unit, a control unit, a fan unit and a valve unit; wherein: the visual unit is responsible for monitoring the smoke dust amount of a first outlet tap and a second outlet tap, and outputs different state signals to the control unit according to the smoke dust amount; the control unit receives the state signals and controls the valve unit and the fan unit to act in response; the valve unit is divided into two parts, namely a valve one (1) and a valve two (2) arranged in the dust removal pipelines of the first outlet tap (3) and the second outlet tap (4); when the first outlet tap (3) and / or the second outlet tap (4) is in a non-tapping stage, the control unit controls the opening degree of the valve one (1) and / or the valve two (2) to decrease; when the first outlet tap (1) and / or the second outlet tap (2) is in a tapping stage, the control unit controls the opening degree of the valve one (1) and / or the valve two (2) to increase; the fan unit comprises a fan (5) which is connected in series in the pipeline between the valve one (1) and the valve two (2); the power of the fan (5) is concentrated at the valve one (1) or the valve two (2) with more smoke dust, so that the utilization rate of the fan is improved, and the control unit controls different rotating speed gears of the fan (5) through a frequency converter to match different smoke dust amounts; The fan (5) is installed behind the intersection point of the dust removal pipelines of the first outlet tap (3) and the second outlet tap (4), directly acts on the dust removal pipelines of the outlet taps, generates airflow, sucks the smoke dust in the dust removal pipelines of the outlet taps and transports the smoke dust to a dust removal equipment; The fan has four basic gears in the automatic mode, namely a low rotating speed gear, a medium rotating speed gear, a high rotating speed gear and an ultrahigh rotating speed gear; when needed, a fixed preset value can be added to the basic gears; The valve one (1) and the valve two (2) are main control valves, and the valve one (1) and the valve two (2) are connected with eight subordinate valves at corresponding positions of the first outlet tap (3) and the second outlet tap (4) respectively.

2. The dual-tap hole blast furnace tap hole area dust removal and power saving control system according to claim 1, characterized in that: The visual unit comprises a camera one (6) and a camera two (7) which are installed at the first outlet tap (3) and the second outlet tap (4) respectively and are used for monitoring the smoke dust amounts of the first outlet tap (3) and the second outlet tap (4) respectively.

3. The dual-tap hole blast furnace tap hole area dust removal and power saving control system according to claim 2, characterized in that: The visual unit comprises a computer (8) which is used for receiving image signals of the camera one (6) and the camera two (7), inputting the signals to a python program for image processing and transmitting state information obtained after processing to the control unit.

4. The dedusting and power saving control system for a double-tap hole blast furnace tapping area according to claim 1 or 2, characterized in that: The control unit comprises a PLC controller (9) and an HMI man-machine interactive interface (10) and has two operation modes, namely an automatic mode and a manual mode.

5. A double-outlet-tap blast furnace tap hole dust removal and power saving control method, which adopts the double-outlet-tap blast furnace tap hole dust removal and power saving control system according to any one of claims 1-4, and is characterized in that: (1) the states of the first outlet tap (3) and the second outlet tap (4) are set as a tapping stage and a non-tapping stage; ①The tapping stage includes two periods: the state of a large amount of smoke dust during tapping is divided into a tapping smoke dust period, and the fan (5) is in a high speed gear or an ultra-high speed gear during this period; the state of a small amount of smoke dust during tapping is divided into a tapping stable period, and the fan (5) is in a medium speed gear during this period; ②The time period from the end of tapping to the start of tapping in the next cycle is divided into a non-tapping stage, and the fan (5) is in a low speed gear during this period; (2) The automatic tapping detection function detects the state of the No. 1 tapping hole (3) and the No. 2 tapping hole (4) through automatic detection, and the specific steps are as follows: The visual unit is responsible for monitoring the amount of smoke dust of the No. 1 tapping hole (3) and the No. 2 tapping hole (4), and transmitting the information to the control unit, and the control unit processes the signal of the visual unit to judge the state of the No. 1 tapping hole (3) and the No. 2 tapping hole (4); (3) The opening degree of valve one and valve two and the gear of the fan are controlled by the control unit, and the specific steps are as follows: ①When the visual unit detects that the No. 1 tapping hole (3) is in the tapping stage and the No. 2 tapping hole (4) is in the non-tapping stage, the control unit adjusts the opening degree of valve one (1) to a high conduction state, the opening degree of valve two (2) to a low conduction state, and the power of the fan (5) is concentrated in valve one (1) where the No. 1 tapping hole (3) is located, and at the same time, the speed of the fan is adjusted to the corresponding gear according to the amount of smoke dust at the No. 1 tapping hole (3); ②When the visual unit detects that the No. 1 tapping hole (3) is in the non-tapping stage and the No. 2 tapping hole (4) is in the tapping stage, the control unit adjusts the opening degree of valve one (1) to a low conduction state, the opening degree of valve two (2) to a high conduction state, and the power of the fan (5) is concentrated in valve two (2) where the No. 2 tapping hole (4) is located, and at the same time, the speed of the fan is adjusted to the corresponding gear according to the amount of smoke dust at the No. 2 tapping hole (4); ③When the visual unit detects that the No. 1 tapping hole (3) and the No. 2 tapping hole (4) are both in the tapping stage, the opening degree of valve one (1) and valve two (2) is adjusted to a high conduction state, and the speed of the fan (5) is controlled to keep in the corresponding gear until the No. 1 tapping hole (3) and the No. 2 tapping hole (4) do not meet the conditions of the tapping stage; ④When the visual unit detects that the No. 1 tapping hole (3) and the No. 2 tapping hole (4) are both in the non-tapping stage, the opening degree of valve one (1) and valve two (2) is adjusted to a low conduction state, and the speed of the fan (5) is controlled to keep in the corresponding gear; The automatic tapping detection function is realized by a deep learning recognition method or a marker recognition method: The deep learning recognition method is: by collecting tapping images of the tapping holes in the three stages, a data set is constructed and a model is trained, which can accurately identify the state of the No. 1 tapping hole and the No. 2 tapping hole; The marker recognition method is: a marker plate is installed behind the smoke dust of the No. 1 tapping hole and the No. 2 tapping hole, the size of the smoke dust is judged by recognizing the green area of the marker plate through OpenCV, and the state of the No. 1 tapping hole and the No. 2 tapping hole is accurately identified; The deep learning identification method and the marker identification method can be used alone or together to identify the amount of smoke, and the identification result of the larger amount of smoke is used as the reference; On the basis of the deep learning identification method and the marker identification method identifying the state of the No. 1 and No. 2 tapping holes, a separate neural network model is added to identify the worker operation; If the visual unit detects that there is worker operation at the No. 1 and No. 2 tapping holes, the fan speed will be increased by a preset value on the basis of the current fan speed; when the worker operation is completed and the worker leaves the site, the fan speed will return to the basic speed; When the visual unit determines that the current state is the non-tapping stage and there is no worker operation, first, the OpenCV is used to identify the position of the marker plate and recalibrate the position, and the color and brightness of the marker plate are identified; if it is lower than a preset value, the HMI indicator light is used to prompt the operator to wipe, and it is suggested that the worker switches to only use the neural network method for judgment, or temporarily adjusts to the manual mode.

Images