Dynamic compensation type sintering machine wind tunnel optimization system and method, terminal and medium

Abstract

The invention belongs to the technical field of sintering machine wind tunnel processing, and particularly discloses and provides a dynamic compensation type sintering machine wind tunnel optimization system and method, a terminal and a medium. Intelligent decision making is achieved in combination with defect map matching, the loading unit is controlled to conduct precise feeding, and the wind tunnel area is effectively blocked. A multi-modal image fusion and image segmentation algorithm is adopted, so that the accuracy and robustness of wind tunnel identification are improved; combining the speed of the sintering pallet to predict a wind tunnel trajectory, positioning a charging unit in advance, and realizing efficient dynamic compensation. The system has closed-loop control capability, can significantly improve the stability and automation level of the sintering process, reduces material waste and manual intervention cost, and is suitable for wind tunnel treatment scenes under complex working conditions.

Description

Technical Field

[0001] The present invention belongs to the technical field of sintering machine wind tunnel treatment, and particularly relates to a dynamic compensation type sintering machine wind tunnel optimization system, method, terminal and medium. Background Art

[0002] As a key pretreatment link in iron and steel production, the uniformity and integrity of the sintering layer in the sintering process directly affect the efficiency and product quality of subsequent blast furnace ironmaking. During the sintering process, fluctuations in the negative pressure in the wind box are likely to cause local air flow channels to form on the surface of the sintering layer, commonly known as "wind tunnels". The generation of wind tunnels will cause some sintered materials not to be fully sintered, resulting in a decrease in product strength, a reduction in fuel utilization rate, and even abnormal shutdown of equipment, seriously affecting the stability of production and economic benefits.

[0003] At present, traditional wind tunnel detection and control methods mainly rely on manual inspections and empirical adjustments, lacking real-time, accurate monitoring and rapid response capabilities, resulting in a significant lag in the identification and blocking of wind tunnels. In addition, the detection of wind tunnels in existing technologies mostly relies on a single sensing signal, such as the negative pressure value, which is difficult to comprehensively reflect the spatial distribution and dynamic changes of wind tunnels, and does not fully combine image information for multimodal fusion analysis, with limited identification accuracy.

[0004] On the other hand, the control strategies for wind tunnel blocking are mostly fixed patterns, lacking dynamic prediction and intelligent matching of the size, position and movement trajectory of wind tunnels, resulting in insufficient feeding accuracy, serious material waste, and inability to effectively adapt to the complex and changeable working conditions of the sintering layer. At the same time, the negative pressure thresholds in existing technologies are mostly fixed settings, and are not dynamically adjusted in combination with key parameters such as waste gas composition, making it difficult to distinguish process fluctuations from wind tunnel abnormalities, affecting the accuracy and stability of the wind tunnel triggering system. Summary of the Invention

[0005] In view of the problems in the prior art, the present invention provides a dynamic compensation type sintering machine wind tunnel optimization system, method, terminal and medium, which solves the problem that the detection of wind tunnels in the prior art mostly relies on a single sensing signal, is difficult to comprehensively reflect the spatial distribution and dynamic changes of wind tunnels, and has limited identification accuracy.

[0006] The technical solutions adopted by the present invention are as follows: In a first aspect, the present application provides a dynamic compensation type sintering machine wind tunnel optimization system, which includes: A mounting frame, on which two mutually parallel guide rails are installed. An image acquisition unit is slidably arranged on one of the guide rails, and the image acquisition unit is used to photograph the sintering layer. A loading unit is arranged on the other guide rail, and the loading unit is used to pour sintered materials into the sintering layer.

[0007] Preferably, the loading unit includes a servo motor and a tipping bucket, and the output shaft of the servo motor abuts against one side of the tipping bucket.

[0008] Preferably, the mounting frame includes two vertically arranged brackets, a cross beam is arranged at the upper ends of the brackets, and both ends of the cross beam are fixedly connected to the upper ends of the two brackets respectively, and the guide rail is arranged on the cross beam.

[0009] In a second aspect, the present application provides a dynamic compensation type sintering machine wind tunnel optimization method, including the following steps: Step S1: Real-time collect the thermal imaging image and visible light image on the surface of the sintering layer, and generate a sintering layer distribution map after preprocessing and fusion processing; Step S2: Obtain the wind tunnel movement speed, identify the wind tunnel area formed by the sintering layer based on the map, and extract the size and position of the wind tunnel area; Step S3: According to the size information of the wind tunnel area extracted in Step S2, perform defect comparison in the preset optimization map, and select the corresponding tipping data; Step S4: Control the loading unit to move above the wind tunnel area and perform tipping; Step S5: Monitor the negative pressure signal. When the negative pressure signal reaches above the threshold, the wind tunnel optimization is completed. Otherwise, an alarm is given, and the operation process and data are recorded.

[0010] Preferably, Step S1 includes the following steps: Step S1-1: Real-time collect the thermal imaging image and visible light image on the surface of the sintering layer; Step S1-2: Perform image processing and two-dimensional wavelet decomposition on the thermal imaging image and the visible light image respectively, and extract the low-frequency component , and the high-frequency component , ; Step 1-3: Select the one with a higher average gray level as the main structure to retain;

[0011]

[0012] Among them, is the image pixel coordinate, is the average gray level value of the thermal imaging area, the average gray level value of the visible light area; Step 1-4: Reconstruct the fused low-frequency and high-frequency subbands and into a sintering layer distribution map through inverse wavelet transform .

[0013] Preferably, in step S1-2, the image is subjected to Gaussian filtering before fusion, and the filtering function G(x,y) is:

[0014] where σ is the standard deviation of the Gaussian kernel.

[0015] Preferably, in step S2, the fused image is input into the image recognition module to extract the wind tunnel area. The size D of the wind tunnel area is calculated according to the coordinates of the recognized image bounding box, specifically:

[0016] where and and and are the maximum and minimum coordinates of the wind tunnel image area in the horizontal and vertical directions respectively; The image segmentation algorithm is used to represent the wind tunnel area as a mask function ; where indicates that the point belongs to the wind tunnel area, indicates that the point does not belong to the wind tunnel area; The wind tunnel area A can be calculated according to the following formula:

[0017] where s is the actual physical length represented by a single pixel, is the area corresponding to a single pixel, , .

[0018] Preferably, in step S3, the extracted wind tunnel size is compared with the preset defect atlas, and the Euclidean distance is calculated to determine the historical defect form with the closest match. The distance formula is as follows:

[0019] where and are the j-th size and area data in the optimized atlas respectively.

[0020] Thirdly, the present application provides a terminal, including: A memory for storing the sintering machine wind tunnel processing simulation program; A processor for executing the steps of the sintering machine wind tunnel processing method as described in the second aspect when implementing the sintering machine wind tunnel processing system.

[0021] Fourthly, the present application provides a computer-readable storage medium storing computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the sintering machine wind tunnel processing method as described in the second aspect.

[0022] As can be seen from the above technical solutions, the advantages of the present invention are as follows: (1) This method realizes a closed-loop control logic based on negative pressure signal triggering, image fusion recognition, trajectory prediction, precise feeding, and feedback verification, can automatically detect and effectively block the wind tunnel area during the sintering process, significantly reduce the risk of manual processing, and improve production continuity and energy consumption stability.

[0023] (2) By introducing an oxygen content factor through a dynamically calculated negative pressure threshold, it effectively distinguishes process fluctuations from wind tunnel abnormalities, improves the accuracy of the trigger recognition system, avoids misoperations, and enhances the pertinence and intelligence of recognition responses; by extracting key parameters such as the size, position, area, and movement trajectory of the wind tunnel, it provides a data basis for wind tunnel morphology modeling and precise feeding; the trajectory prediction ability enables the loading unit to achieve dynamic alignment, improve the blocking accuracy, and reduce material waste.

[0024] (3) Adopting a multimodal image fusion method to complementarily fuse the thermal imaging map and the visible light map, retaining the structural boundary and temperature anomaly information, making the generated sintering layer map have stronger feature expression ability, improving the clarity and reliability of wind tunnel recognition; introducing Gaussian filtering processing before image fusion can effectively remove image noise, enhance the clarity of the edge contour, provide a better image basis for subsequent wind tunnel boundary extraction, and enhance the system's recognition ability for small-scale wind tunnels.

[0025] (4) Comparing the identified wind tunnel morphology with the preset defect map to realize the selection of an intelligent feeding strategy based on historical data, which helps to form an integrated linkage of "recognition - decision-making", improve the adaptability and success rate of the blocking behavior; realizing the early positioning of the loading unit according to the trolley running speed and system response delay, effectively avoiding misaligned feeding caused by lag, ensuring that the material accurately falls into the wind tunnel area, and further enhancing the timeliness and robustness of automatic control. Description of the Drawings

[0026] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0027] Figure 1 It is the method flow chart of the embodiment of the specific implementation manner of the present invention; Figure 2 Schematic diagram of the system structure of the embodiment of the specific implementation manner of the present invention Figure 1 ; Figure 3 Schematic diagram of the system structure of the embodiment of the specific implementation manner of the present invention Figure 2 。

[0028] In the figure: 1, support; 2, cross beam; 3, loading unit; 4, image acquisition unit; 5, guide rail; 6, wind tunnel. Specific implementation manner

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0030] Please refer to Figure 1 As shown, the present invention provides a dynamic compensation type sintering machine wind tunnel optimization method, including the following steps: Step S0, obtain the negative pressure signal inside the sintering machine air box in real time during the sintering process. When the negative pressure signal is lower than the threshold, execute step S2; The negative pressure of the sintering machine air box reflects the uniformity and smoothness of the air flow through the sintering layer. When the grate bars are loose, deformed or fallen off to form local channels, it will cause abnormal high-speed air flow channels in some areas, that is, the wind tunnel phenomenon, and then cause a significant decrease in the negative pressure value at that place. Therefore, by collecting and analyzing the negative pressure data in the air box in real time, it is possible to effectively judge whether a wind tunnel appears.

[0031] A plurality of pressure detection devices are arranged along the length direction of the sintering machine air box. The system collects the negative pressure data of each detection point at regular intervals and takes the average value as the judgment basis. When the system detects that the average negative pressure value is lower than the preset safety threshold and lasts for the set time, it automatically determines that there may be a wind tunnel phenomenon in the sintering layer and immediately starts the image recognition process.

[0032] Step S1, collect the thermal imaging image and visible light image on the surface of the sintering layer in real time, and generate a sintering layer distribution map after preprocessing and fusion processing; The thermal imaging image can reflect the temperature distribution on the surface of the sintering layer and can quickly locate the temperature abnormal area; the visible light image is used to identify the structure contour of the sintering layer and the grate bar gap. By fusing the two types of images, a distribution map with both thermal information and structural information can be obtained, which helps to more accurately identify the wind tunnel area.

[0033] After the system enters the image acquisition mode, the infrared thermal imaging device and the visible light imaging device installed above the sintering trolley are respectively started to collect the image data of the upper surface of the sintering layer at the same time. The system performs unified preprocessing operations on the two types of images, including denoising, alignment calibration, edge enhancement, etc. Subsequently, the two images are fused according to preset rules to generate a fused sintering layer atlas for subsequent analysis of the position and shape of the wind tunnel area.

[0034] Step S2: Obtain the movement speed of the sintering layer, identify the wind tunnel area formed by the sintering layer and the movement speed data of the sintering layer based on the atlas, extract the size, position and movement trajectory characteristics of the wind tunnel area, and calculate the predicted trajectory of the wind tunnel area based on the movement trajectory of the wind tunnel area and the movement speed of the sintering layer; The sintering layer continues to move forward driven by the trolley, and the wind tunnel area has spatial mobility. In order to achieve precise feeding and plugging, the system needs to identify the position, size and movement trajectory of the wind tunnel over time, and perform prediction compensation in combination with the trolley movement speed, so as to ensure that the feeding unit can be accurately aligned.

[0035] The system first reads the trolley walking speed data and combines it with the atlas recognition result. Through the intelligent image analysis module, it identifies the areas with sudden brightness change, edge break or local low temperature anomaly in the atlas and determines them as wind tunnels. The system further extracts the boundary contour of this area in the image and calculates its position, length, width and shape characteristics. Subsequently, according to the trolley movement direction and speed information, it predicts the position change of the wind tunnel area in the next time period and generates dynamic trajectory information for the reference of the control module.

[0036] Step S3: According to the size information of the wind tunnel area extracted in Step S2, perform defect comparison in the preset optimized atlas and select the corresponding tipping data; The wind tunnel morphology has diversity due to differences in position, size and structure. The system selects a matching atlas template by comparing the actual morphology of the wind tunnel with the typical defect samples in the historical atlas database, and extracts the preset optimal tipping scheme from it to ensure that corresponding feeding control strategies can be obtained for different types of wind tunnels.

[0037] The system internally stores multiple historical typical wind tunnel morphology atlases and corresponding tipping parameter data. When the current wind tunnel is identified, the system automatically compares the size and contour morphology of the wind tunnel with all templates in the atlas library, and selects the closest sample through similarity ranking. The system determines the quantity, dumping angle and duration of the feeding of the feeding unit according to the feeding strategy set in this sample, and sends an instruction to prepare to enter the feeding link.

[0038] Step S4: Control the feeding unit to move above the predicted trajectory of the wind tunnel area and perform tipping; The wind tunnel area is a moving target, and the loading unit needs to reach above its predicted position in advance and complete fixed-point feeding. Through trajectory prediction and dynamic positioning control, the precise synchronization of the movement trajectories of the loading mechanism and the wind tunnel is achieved, ensuring that the feeding landing point coincides with the wind tunnel.

[0039] The system calculates the position change trend of the wind tunnel in the next few seconds through a trajectory prediction model, and determines a feeding target position in combination with the response time of the feeding mechanism and the trolley speed. After controlling the loading unit to move along the transverse guide rail to above the target position, the feeding mechanism is started. The feeding mechanism uses an electric push rod to control the tipping angle and pours the pre-loaded sintered ore material into the wind tunnel area within a set time to block the air flow channel.

[0040] Step S5: Monitor the negative pressure signal. When the negative pressure signal reaches above the threshold, the wind tunnel optimization is completed; otherwise, an alarm is issued, and the operation process and data are recorded. After the feeding is completed, it is necessary to detect the change of the wind box negative pressure again to judge whether the blockage is successful. If the negative pressure returns to normal, it means that the wind tunnel is effectively filled; if the negative pressure does not recover, it may be that the blockage fails or the position of the wind tunnel is misjudged, and an alarm prompt should be issued, and all key data should be recorded for subsequent analysis and optimization of the control logic.

[0041] After the feeding is over, the system continuously collects the pressure data in the wind box and judges whether it returns above the safety threshold. If the negative pressure returns to normal, the system automatically records the time, position, feeding parameters, blockage effect and image data of this wind tunnel treatment and archives them in the database. If the negative pressure still does not recover, the system automatically alarms, prompts manual confirmation, and also records the current processing process as an abnormal case for the subsequent self-learning and optimization of the system.

[0042] In some embodiments, in step S0, the negative pressure signal is the average value P of the detection values of multiple pressure detection units, and the negative pressure threshold is dynamically calculated according to the following formula:

[0043] where is the historical negative pressure average value under normal working conditions, is the change value of the waste gas oxygen content, and k is an adjustment coefficient.

[0044] In some embodiments, step S1 includes the following steps: Step S1-1: Real-time collect the thermal imaging image and visible light image of the sintering layer surface; Step S1-2: Perform image processing and two-dimensional wavelet decomposition on the thermal imaging image and the visible light image respectively, and extract the low-frequency components , and the high-frequency components , ; Step 1-3: Select the one with a higher average grayscale as the main structure and retain it;

[0045]

[0046] Among them, is the image pixel coordinate, is the average grayscale value of the thermal imaging area, the average grayscale value of the visible light area; Step 1-4: Reconstruct the fused low-frequency and high-frequency sub-bands and into a sintering layer distribution map through inverse wavelet transform .

[0047] In some embodiments, in step S1-2, before fusion, the image is subjected to Gaussian filtering, and the filtering function G(x,y) is:

[0048] Among them, σ is the standard deviation of the Gaussian kernel.

[0049] In some embodiments, in step S2, the fused image is input into an image recognition module to extract the wind tunnel area. The size D of the wind tunnel area is calculated according to the recognized image bounding box coordinates, specifically:

[0050] Among them, , , , are the maximum and minimum coordinates of the wind tunnel image area in the horizontal and vertical directions respectively; Adopt an image segmentation algorithm to represent the wind tunnel area as a mask function ; Among them indicates that this point belongs to the wind tunnel area, indicates that this point does not belong to the wind tunnel area; The wind tunnel area A can be calculated according to the following formula:

[0051] Among them, s is the actual physical length represented by a single pixel, is the area corresponding to a single pixel, , ; Combined with the sintering trolley speed v, predict the change trajectory X(t) of the wind tunnel position over time, and the expression is:

[0052] Among them, is the initial recognition position, and t is the time.

[0053] In some embodiments, in step S3, the extracted wind tunnel dimensions are compared with a preset defect atlas, and the most closely matched historical defect form is judged by calculating the Euclidean distance. The distance formula is as follows:

[0054] Among them, and are the j-th dimension and area data in the optimized atlas respectively.

[0055] In some embodiments, the charging unit is controlled to perform early movement positioning according to the trolley speed v and the predicted trajectory X(t) of the wind tunnel. The positioning target position is:

[0056] Among them, is the system positioning compensation time.

[0057] In some embodiments, the present application provides a dynamic compensation type sintering machine wind tunnel optimization system, which includes: A negative pressure detection module, which is used to be arranged inside the wind box, collect negative pressure data during the sintering process in real time, and output an average negative pressure value; This module is used to monitor the negative pressure change in the wind box in real time and judge whether the ventilation of the sintering layer is normal. The formation of the wind tunnel usually causes a rapid decrease in local negative pressure. Therefore, abnormal areas can be detected in time through multi-point pressure data; Pressure sensors are installed at multiple positions of the wind box under the sintering machine trolley, for example, one group is set every one meter. The system controller reads the pressure data of all sensors at regular intervals, performs averaging processing, and forms the current negative pressure level data. This module is linked with the control logic. When the average value is lower than the preset safety threshold, the image detection process is automatically started.

[0058] An image acquisition module, including an infrared thermal imager and a visible light camera, is used to simultaneously obtain infrared images and visible light images of the surface of the sintering layer; This module is used to obtain visual information and temperature information on the surface of the sintering layer, which is the basis for identifying the shape of the wind tunnel. The thermal imager can display areas with temperature differences, while the visible light image can present physical contours and surface structures; An infrared thermal imaging camera and a visible light camera are installed above the running area of the sintering trolley. The two cameras synchronously collect images of the same area and maintain frame consistency through a time synchronization mechanism. The image data is directly transmitted to the image processing unit at the back end and enters the fusion and recognition process.

[0059] The image fusion and recognition module, including an image preprocessing unit, a wavelet transform processing unit, and an image segmentation unit, is used to filter, extract features, and perform image fusion processing on the acquired images, identify the wind tunnel area, and obtain the size, position, and movement trajectory of the wind tunnel area; This module performs clarification processing, fusion integration on the acquired images, and identifies the shape, range, and position of the wind tunnel area as the basis for subsequent control; The system first performs noise filtering and edge enhancement processing on the thermal imaging map and the visible light map respectively. Then, the two types of images are uniformly registered, and the two types of information are merged through an image fusion algorithm to generate a composite map with both structural and temperature information. The system automatically analyzes the areas with large temperature differences or structural fractures in the map, identifies them as the wind tunnel area, and outputs its location, occupied area, and consistency with the advancing direction of the sintering layer.

[0060] The feeding control module, including a loading unit, a driving mechanism, and a positioning control unit, is used to perform lateral movement positioning of the loading unit according to the predicted trajectory of the wind tunnel and execute the feeding operation; This module undertakes the execution action, that is, accurately putting specific materials into the wind tunnel according to the predicted position of the wind tunnel to achieve rapid plugging; The loading unit is installed on the lateral guide rail above the trolley and has the ability to move along the guide rail. The drive system is controlled by a servo motor and moves the loading unit above the target area according to the predicted position. The controller controls the tipping angle and opening and closing time according to the feeding parameters to complete the plugging action. After the feeding is completed, the device automatically resets and prepares for the next action.

[0061] The trajectory prediction and control module is used to predict the target position of the wind tunnel according to the running speed of the sintering trolley and the system positioning compensation time, and send a positioning instruction to the feeding control module; Since the sintering trolley is moving continuously, the position of the wind tunnel is not a fixed point. It is necessary to predict the future position of the wind tunnel in combination with the current running speed to ensure the accurate positioning of the loading unit; The system obtains the real-time speed by reading the signal of the trolley running encoder, and then combines the current position of the wind tunnel and the system response time to calculate the expected position of the wind tunnel in the future. The prediction result is transmitted as the target input to the positioning control system to guide the loading unit to align with the target point in advance.

[0062] The map matching unit is used to compare the identified wind tunnel shape parameters with the preset defect maps and select the corresponding feeding parameters for optimized control; This module is used to compare the currently identified wind tunnel characteristics with historical samples, so as to select the best feeding strategy and achieve personalized control of "applying materials according to the hole"; The system pre-stores multiple typical wind tunnel morphology samples and corresponding plugging strategies. After the current wind tunnel is identified, the system retrieves all the sample atlases from the database, calculates the similarity between the current wind tunnel and the samples, selects the most matching one, and reads the set feeding weight, angle, and feeding time as the control parameters for this plugging operation.

[0063] The system control and recording module, including a programmable logic controller, an operation interface, and a database system, is used to control the operation process of the entire system, process alarm information, and automatically record operation logs such as image data, feeding data, and negative pressure changes.

[0064] This module is responsible for managing the coordinated operation of the entire system, processing various data flows, command distributions, and information storage, and is the "central nerve" of the entire set of systems; The system uses an industrial-grade programmable logic controller as the core control unit to connect all execution units such as sensors, cameras, and drivers. Parameters can be set, manual intervention can be performed, or the operating status can be viewed through the operation interface. The controller classifies and organizes each wind tunnel identification, prediction, feeding, and feedback information and stores it in the database. The database supports queries according to dimensions such as time, area, and success rate, facilitating subsequent maintenance and continuous optimization of the system.

[0065] Please refer to Figure 2 and Figure 3 As shown, in some embodiments of the present invention, during the operation of the dynamic compensation type sintering machine wind tunnel optimization system, first, through the pressure detection module set at multiple positions of the sintering machine air box, the negative pressure data below the sintering area is collected in real time, and the values of multiple detection points are averaged to determine whether there is an abnormality in the current air flow. When the system continuously detects that the average negative pressure is lower than the preset safety threshold value and lasts for more than the set time threshold, the image acquisition process is automatically triggered.

[0066] Subsequently, the image acquisition module is activated. The image acquisition module is the image acquisition unit 4. The image acquisition unit 4 and the loading unit 3 are respectively arranged on two guide rails 5 and can move left and right. The guide rails 5 are installed on the cross beam 2, and the cross beam 2 is arranged on the bracket 1. The infrared thermal imaging device and the visible light camera installed above the sintering trolley synchronously acquire images of the sintering layer surface. The acquired image signals are transmitted to the image fusion and recognition module through the data interface for processing. This module first preprocesses the images, including operations such as filtering and noise reduction, edge sharpening, and geometric registration, to improve the image clarity and alignment. The preprocessed images are synthesized through a fusion algorithm to generate a composite atlas containing temperature distribution information and structural boundary features. The system automatically identifies the suspected wind tunnel 6 area from this atlas and extracts key parameters such as its position, range, shape, and relative movement trend.

[0067] After the recognition is completed, the map matching unit is activated. It compares and analyzes the currently recognized wind tunnel form parameters with the typical defect maps preset in the system, selects the historical wind tunnel model with the highest matching degree, and extracts the corresponding feeding control parameters, including information such as the required material quantity, feeding angle, and duration. At the same time, the trajectory prediction and control module receives the initial position of the wind tunnel obtained by recognition and the current running speed of the sintering trolley, calculates the target position that the wind tunnel will reach in the future by combining the system response time, and sends it as the positioning target to the feeding control module.

[0068] After receiving the prediction result, the feeding control module drives the loading unit installed on the transverse track above the trolley to move to the predicted target position. The loading unit is internally provided with multiple storage chambers and a controllable tipping device. After reaching the target position, the system controls the tipping angle and duration according to the feeding parameters, and precisely discharges the preloaded large-particle sinter ore into the wind tunnel area to effectively block the high-speed air flow channel.

[0069] After the blocking operation is completed, the system control module collects the data of the negative pressure detection module again and determines whether the current negative pressure has returned to the normal range. If the detection result shows that the negative pressure has returned above the safe level, the system confirms that the wind tunnel blocking is successful this time, and uploads information such as the recognition data, image records, feeding parameters, and detection feedback to the database to complete the whole process record; if the detection result has not recovered, the system will automatically issue an alarm prompt, and at the same time retain the current recognition and operation data as an abnormal sample for subsequent manual intervention and system learning and optimization.

[0070] In some embodiments, the present application provides a terminal, including: A memory for storing the sintering machine wind tunnel processing simulation program; A processor for implementing the steps of the sintering machine wind tunnel processing method when executing the sintering machine wind tunnel processing system.

[0071] In some embodiments, the present application provides a computer-readable storage medium, and the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the sintering machine wind tunnel processing method.

[0072] The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of this technology make various modifications or supplements to the described specific embodiments or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all belong to the protection scope of the present invention.

Claims

1. A dynamic compensation type sintering machine wind tunnel optimization system, characterized in that, It includes a mounting frame, on which two mutually parallel guide rails (5) are mounted. An image acquisition unit (4) is slidably arranged on one of the guide rails (5), and the image acquisition unit (4) is used to photograph the sintered layer. A loading unit (3) is arranged on the other guide rail (5), and the loading unit (3) is used to pour the sintering material into the sintered layer.

2. The dynamic compensation type sintering machine wind tunnel optimization system according to claim 1, characterized in that, The loading unit (3) includes a servo motor and a tipping bucket, and the output shaft of the servo motor abuts against one side of the tipping bucket.

3. The dynamic compensation type sintering machine air duct optimization system according to claim 1, wherein The mounting frame includes two vertically arranged brackets (1). A cross beam (2) is arranged at the upper end of the bracket (1), and both ends of the cross beam (2) are fixedly connected to the upper ends of the two brackets (1) respectively. The guide rail (5) is arranged on the cross beam (2).

4. A dynamic compensation type sintering machine air duct optimization method, characterized in that, The method includes: Step S1: Real-time collect the thermal imaging image and visible light image of the surface of the sintered layer, and generate a sintered layer distribution map after preprocessing and fusion processing; Step S2: Obtain the movement speed of the wind tunnel, identify the wind tunnel area formed by the sintered layer based on the map, and extract the size and position of the wind tunnel area; Step S3: According to the size information of the wind tunnel area extracted in Step S2, conduct defect comparison in a preset optimization map, and select the corresponding pouring data; Step S4: Control the loading unit to move above the wind tunnel area and pour the material; Step S5: Monitor the negative pressure signal. When the negative pressure signal reaches above the threshold, complete the wind tunnel optimization. Otherwise, give an alarm and record the operation process and data.

5. The dynamic compensation type sintering machine air duct optimization method according to claim 1, characterized in that Step S1 includes the following steps: Step S1-1: Real-time collect the thermal imaging image and visible light image of the surface of the sintered layer; Step S1-2: Perform image processing and two-dimensional wavelet decomposition on the thermal imaging image and the visible light image respectively, extract the low-frequency components and the high-frequency components and ; Step 1-3: Select the one with a higher average gray level as the main structure and retain it; Among them, is the image pixel coordinate, is the average gray value of the thermal imaging area, and the average gray value of the visible light area; Steps 1-4, reconstruct the fused low-frequency and high-frequency subbands and into a sintered layer distribution map through inverse wavelet transform .

6. The dynamic compensation type sintering machine air duct optimization method according to claim 5, characterized in that In Step S1-2, perform Gaussian filtering on the image before fusion, and the filtering function G(x, y) is: where σ is the standard deviation of the Gaussian kernel.

7. The dynamic compensation type sintering machine air duct optimization method according to claim 4, wherein In Step S2, the fused image is input into the image recognition module to extract the wind tunnel area. The size D of the wind tunnel area is calculated according to the coordinates of the recognized image bounding box, specifically: Among them, , , , are the maximum and minimum coordinates of the wind tunnel image area in the horizontal and vertical directions, respectively; Using an image segmentation algorithm, represent the wind tunnel area as a mask function ; Among them indicates that this point belongs to the wind tunnel area indicates that this point does not belong to the wind tunnel area The wind tunnel area A can be calculated according to the following formula: where s is the actual physical length represented by a single pixel, is the area corresponding to a single pixel, , .

8. The dynamic compensation type sintering machine air duct optimization method according to claim 7, characterized in that In Step S3, the extracted wind tunnel size is compared with the preset defect map, and the Euclidean distance is calculated to judge the historical defect form with the closest match. The distance formula is as follows: Among them, and are the j-th dimension and area data in the optimized map, respectively.

9. A terminal, characterized in that, It includes: A memory, used to store the sintering machine wind tunnel processing simulation program; A processor, used to execute the steps of the sintering machine wind tunnel processing method as described in Claim 1 when implementing the sintering machine wind tunnel processing system.

10. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the sintering machine wind tunnel processing method as described in Claim 1.

Images