Converter waste heat boiler flue intelligent detection method and system

Abstract

The invention discloses an intelligent detection method and system for a converter waste heat boiler flue, and belongs to the technical field of intelligent detection of industrial equipment. According to the system, infrared imaging and visible light acquisition equipment is carried by means of an unmanned aerial vehicle, full-coverage scanning is carried out on the flue water wall under the specific working condition of a converter, and intelligent recognition and positioning of cracks, abrasion and other damage are achieved by combining three-dimensional modeling, infrared thermal imaging analysis and historical data dynamic comparison. The method comprises the steps of flight of a preset route, multi-modal data acquisition, intelligent data processing and damage prediction, and establishment of a detection database to realize damage trend analysis. The method solves the problems of potential safety hazards and efficiency of traditional manual detection, provides accurate data support for preventive maintenance of converter equipment, and has the remarkable advantages of automation, intelligence and high reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of converter flue gas waste heat boiler flue damage detection, and particularly relates to a converter waste heat boiler flue intelligent detection method and system. BACKGROUND

[0002] The converter is the core equipment for realizing steel smelting, and the converter waste heat boiler is an important device for recovering the high-temperature flue gas waste heat generated in the steel smelting process of the converter, and has important significance for improving energy utilization and reducing production cost. The smelting cycle of the converter is about 38 minutes, and the oxygen blowing time is about 15 minutes. The flue gas generated during oxygen blowing has a temperature of about 1500 DEG C, and the flue dust content in the flue gas is as high as 120 g / m 3 The water-cooled wall flue is a main component of the waste heat boiler. As a component directly contacting the high-temperature flue gas, the flue is long-term exposed to a harsh environment of periodic high temperature (200-1500 DEG C), high pressure (circulating water pressure 2.5 Mpa) and high-speed flue dust erosion, and is prone to fatigue cracks and wear thinning damage. If these damages cannot be found and repaired in time, the position of the water-cooled wall pipe with serious cracks and excessive wear will be broken over time, and the high-pressure cooling water will leak. Once this happens, not only the normal operation of the waste heat boiler will be affected, and the waste heat recovery efficiency will be reduced, but also a large amount of flue cooling water will flow into the converter, the water will instantaneously vaporize and explode in the molten steel, endangering the life safety of the operating personnel, causing huge economic losses and adverse social impact. At present, the damage detection of the water-cooled wall flue of the converter waste heat boiler mainly adopts the following methods: During the regular maintenance of the converter, the operator enters the flue from the maintenance opening at the upper part of the flue by means of a basket, and then checks the inner wall of the flue by means of light visual inspection, ultrasonic wave and other tools. This method has obvious defects: first, the inside of the flue is narrow and dark, and the operator has limited vision, so it is easy to miss detection; second, the diameter of the flue is as large as 2.5 m or more, and the length is as long as 30 m or more, so the area to be detected is huge, and manual detection takes too long, which seriously affects production; third, manual detection depends on the experience and responsibility of the detection personnel, so it is easy to miss detection or misdiagnosis, and the detection accuracy and reliability are difficult to guarantee; fourth, the maintenance cycle of the converter is as long as 3-6 months, so the flue damage cannot be found in time.

[0003] Therefore, it is of great practical significance to develop a converter waste heat boiler flue intelligent detection method and system to solve the problems of poor safety, low detection accuracy, low efficiency and inability to dynamically track in the existing detection technology, so as to ensure the stable operation of the converter production, improve the production efficiency and reduce the production cost. SUMMARY

[0004] The present invention aims to solve the problems existing in the existing detection methods for damage to the water-cooled walls of the converter waste heat boiler flue, such as low safety, insufficient detection accuracy, low efficiency, and inability to achieve dynamic tracking detection. It provides an intelligent detection method and system for the converter waste heat boiler flue, which can use drones carrying infrared imaging equipment to perform comprehensive and accurate detection of the flue water-cooled walls without affecting converter production. By analyzing infrared images and comparing them with databases, it can realize dynamic identification and positioning of damage, providing a reliable basis for timely repair.

[0005] In a first aspect, the present invention provides a converter waste heat boiler flue intelligent detection system, comprising: a drone unit, an imaging unit, a control unit, a data processing unit, and a database unit; The drone unit is used to achieve autonomous positioning and navigation in a GPS-free environment within the flue; The imaging unit is installed on the UAV unit and includes a visible light camera and an infrared camera. The visible light camera is used to take ordinary photos of the flue water-cooled wall, and the infrared camera is used to capture infrared thermal images of the flue water-cooled wall. The control unit is used to remotely control the drone unit and the imaging unit; The data processing unit is used to construct a three-dimensional model of the flue water-cooled wall based on the flight trajectory of the drone, ordinary pictures and infrared thermal images, identify abnormal areas by analyzing the temperature of the infrared thermal images, and search for corresponding historical data stored in the database unit through the infrared thermal images and the abnormal areas, so as to determine whether the abnormal areas are potential damage locations and mark them in the three-dimensional model.

[0006] In some examples, the data processing unit includes an image preprocessing module, a three-dimensional modeling module, a temperature analysis module, and a comparative analysis module; The image preprocessing module is used to process ordinary photos and infrared thermal images; The three-dimensional modeling module is used to construct a three-dimensional model of the flue water wall based on the flight trajectory of the UAV, the processed ordinary photos and the infrared thermal images, and map the temperature information of the processed infrared thermal images to the corresponding positions of the three-dimensional model to form a three-dimensional model with temperature information; The temperature analysis module is used to perform point-by-point temperature analysis on the three-dimensional model with temperature information, find a first area where the local temperature is less than a first preset temperature threshold, and mark the identified first area as a potential damage area; The comparative analysis module is used for comparing the current detected infrared thermal image data with the historical infrared thermal image data stored in the database unit, finding out a second region of local temperature drop in the historical infrared thermal image, determining a possible damage position from the first region and the second region, and comprehensively judging in combination with structural features and operation conditions of the flue to finally determine the position needing to be overhauled.

[0007] In some examples, the comparative analysis module is used for calling the latest detection data of the converter waste heat boiler flue water wall from the database unit, including historical infrared thermal image temperature distribution data and a historical three-dimensional model, performing registration on the current detection three-dimensional model and the historical three-dimensional model to ensure spatial position consistency of both, performing accurate alignment on the current detection infrared thermal image and the historical infrared thermal image, calculating temperature difference values of corresponding positions, marking a region as a second region of temperature abnormal drop when the current temperature of the region compared with the historical temperature drops by more than a second preset temperature threshold, performing superposition analysis on the first region and the second region, taking an intersection of the first region and the second region as a possible damage position, and comprehensively judging in combination with structural features and operation conditions of the flue for positions appearing only in the first region or the second region to exclude temperature changes caused by cooling water flow fluctuation or local structural difference, and finally determining the position needing to be overhauled.

[0008] In some examples, the control unit includes a ground controller and a UAV on-board controller, the ground controller is connected with the UAV on-board controller through high-temperature-resistant optical fibers, so that an operator sets flight parameters, flight routes and imaging parameters of the UAV on the ground through the ground controller to monitor flight states and imaging conditions of the UAV in real time.

[0009] In some examples, the data processing unit is further used for generating a detection report, including a three-dimensional model of the flue water wall with damage position marking, a temperature distribution chart of the first region, a comparative analysis chart with historical data, and detailed coordinate and temperature information of the possible damage position.

[0010] In the second aspect, the present application provides a converter waste heat boiler flue intelligent detection method, comprising: The UAV unit is positioned and navigated autonomously in the flue without GPS environment through remote control, and normal photos of the flue water wall and infrared thermal images of the flue water wall are shot; A three-dimensional model of the flue water wall is constructed according to the flight trajectory of the UAV, the normal photos and the infrared thermal images, an abnormal region is identified through analysis of the temperature of the infrared thermal images, corresponding historical data is found out from the historical data stored in the database unit through the infrared thermal images and the abnormal region, so as to judge whether the abnormal region is a potential damage position, and the three-dimensional model is marked.

[0011] In some examples, the three-dimensional model of the flue water-cooled wall is constructed according to the flight trajectory of the unmanned aerial vehicle, the normal picture and the infrared thermal image, an abnormal area is identified by analyzing the temperature of the infrared thermal image, corresponding historical data is found out by searching the historical data stored in the database unit through the infrared thermal image and the abnormal area, whether the abnormal area is a potential damage position is judged, and the three-dimensional model is marked, including: The normal picture and the infrared thermal image are processed, the three-dimensional model of the flue water-cooled wall is constructed according to the flight trajectory of the unmanned aerial vehicle, the processed normal picture and the infrared thermal image, and the temperature information of the processed infrared thermal image is mapped to the corresponding position of the three-dimensional model to form a three-dimensional model with temperature information; The three-dimensional model with temperature information is analyzed point by point, a first area with a local temperature less than a first preset temperature threshold is found, and the identified first area is marked as a potential damage area; The current detected infrared thermal image data is compared with the historical infrared thermal image data stored in the database unit, a second area with a local temperature drop in the historical infrared thermal image is found, a possible damage position is determined from the first area and the second area, and a comprehensive judgment is made in combination with the structural characteristics and operation of the flue to finally determine the position to be overhauled.

[0012] In some examples, the current detected infrared thermal image data is compared with the historical infrared thermal image data stored in the database unit, a second area with a local temperature drop in the historical infrared thermal image is found, a possible damage position is determined from the first area and the second area, and a comprehensive judgment is made in combination with the structural characteristics and operation of the flue to finally determine the position to be overhauled, including: The last detection data of the converter waste heat boiler flue water-cooled wall is called from the database unit, including the temperature distribution data of the historical infrared thermal image and the historical three-dimensional model, the three-dimensional model of this detection is registered with the historical three-dimensional model to ensure that the spatial positions of the two are consistent, the infrared thermal image of this detection is accurately aligned with the historical infrared thermal image, the temperature difference of the corresponding position is calculated, when the temperature of a certain area of this detection decreases by more than a second preset temperature threshold compared with the historical temperature, the area is marked as a second area with an abnormal temperature drop, the first area and the second area are superimposed and analyzed, the intersection of the first area and the second area is taken as a possible damage position, for the position appearing only in the first area or the second area, a comprehensive judgment is made in combination with the structural characteristics and operation of the flue to exclude the temperature change caused by the fluctuation of cooling water flow or local structural difference, and finally the position to be overhauled is determined.

[0013] In some examples, the method further includes: Generate a detection report, including a three-dimensional model of the flue water-cooled wall with damage location markers, a temperature distribution chart of the first area, a comparison analysis chart with historical data, and detailed coordinates and temperature information of possible damage locations.

[0014] In some examples, the method further comprises: Training historical detection data and actual damage conditions through a machine learning model to identify temperature characteristics corresponding to different types of damage, and then identifying potential damage areas through the trained machine learning model on the three-dimensional model with temperature information.

[0015] Overall, compared with the prior art, the above technical solutions conceived by the present application can achieve the following beneficial effects: The present application combines unmanned aerial vehicle technology, infrared imaging technology, data processing technology and database technology to achieve dynamic, efficient and accurate detection of converter waste heat boiler flue water-cooled wall damage. The system and method can operate safely under specific conditions of the converter, fully capture the temperature information of the water-cooled wall, and timely discover potential damage through data analysis and historical comparison, providing a reliable basis for timely maintenance of the equipment, effectively reducing production risk and improving production efficiency, and having wide application prospects and practical value. The specific performance is as follows: 1. Improve detection safety: The present application uses unmanned aerial vehicles to enter the flue for detection under specific conditions of the converter, avoiding manual entry into harsh environments and eliminating safety risks for detection personnel, ensuring personnel safety. 2. Improve detection comprehensiveness and accuracy: The unmanned aerial vehicle flies according to a preset route, combined with a high-definition infrared thermal imager, to achieve comprehensive coverage detection of the flue water-cooled wall, accurately capture local temperature changes, improve damage detection accuracy, and reduce missed and false detections. 3. Realize dynamic tracking detection: By establishing a database to store detection data over time, the damage to the flue water-cooled wall can be dynamically compared and analyzed to discover damage development and changes in a timely manner, providing a basis for preventive maintenance and avoiding sudden failures. 4. Improve production efficiency: The detection process of the present application can be completed in a short time after the converter stops blowing oxygen and out of steel deflection, without affecting the normal production rhythm of the converter; at the same time, accurate detection results can guide maintenance work, shorten maintenance time and improve production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0017] Figure 1 is a schematic diagram of the system provided by the embodiment of the present application; Figure 2 is a schematic diagram of the method provided by the embodiment of the present application; Figure 3 is a schematic diagram of the flight route planning of the unmanned aerial vehicle provided by the embodiment of the present application; In Figure 1, 1. unmanned aerial vehicle unit, 2. imaging unit, 3. control unit, 4. data processing unit, 5. database unit, 6. converter waste heat boiler flue, 7. fixed-point take-off position, 8. converter, 9. converter waste heat boiler flue, 10. planned flight route (round trip). DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0019] In the following description, the specific embodiments of the present application will be described with reference to the steps and symbols executed by one or more computers, unless otherwise specified. Therefore, these steps and operations will be mentioned several times by the computer execution, and the computer execution referred to herein includes the operation of the computer processing unit represented by the electronic signal in a structured form. This operation transforms the data or maintains it at the location in the memory system of the computer, which can reconfigure or otherwise change the operation of the computer in a manner known to those skilled in the art. The data structure maintained by the data is the physical location of the memory, which has specific characteristics defined by the data format. However, the principles of the present application are described in the above description, which does not represent a limitation, and those skilled in the art will understand that the following steps and operations can also be implemented in hardware.

[0020] The term "module" or "unit" used herein can be regarded as a software object executed on the operating system. Different components, modules, engines and services herein can be regarded as implementation objects on the operating system. The apparatus and method herein are preferably implemented in software, and of course can also be implemented in hardware, both of which are within the protection scope of the present application.

[0021] As will be understood by those familiar with the art, the terms "one," "a," and "that" as used herein can mean "one or more" or "at least one," and includes the plural unless specifically stated otherwise. It should also be understood that the term "including" as used herein, means "including, but not limited to." Also, as used herein, the term "connected" or "coupled" can mean either a direct connection or coupling or an indirect connection or coupling between or among two or more elements, via another element. Further, the term "connection" or "coupling" as used herein is not restricted to direct connections or couplings. Also, as used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. The foregoing detailed description has set forth various embodiments of the devices and / or processes via the use of specific terminology. However, embodiments can be implemented with other technologies. The terms "device" and "system" are used generically and each can refer to either a device or a system. Each embodiment of a device or system can be implemented in various ways.

[0022] As shown in Figure 1 The intelligent detection system for the flue of the converter waste heat boiler provided by the embodiment of the present application comprises: The unmanned aerial vehicle unit 1 is used for carrying a common optical camera and an infrared camera. The imaging unit 2 is installed on the unmanned aerial vehicle unit 1 and comprises a high-definition common optical camera and an infrared camera and is used for collecting image data and heat distribution of the inner wall of the flue. The control unit 3 comprises a ground controller and an airborne controller and cooperatively realizes unmanned aerial vehicle task initialization, flight control and image collection instruction issuing. The ground controller and the airborne controller are connected by high-temperature-resistant optical fibers. The data processing unit 4 comprises an image processing module, a three-dimensional modeling module, a temperature analysis module and a comparative analysis module and is used for processing images and identifying damages. The database unit 5 is used for storing detection images, three-dimensional models, historical data and detection reports.

[0023] Further, the unmanned aerial vehicle unit carries a high-precision inertial navigation system (containing a gyroscope and an acceleration sensor) and combines ultrasonic radar SLAM technology to realize autonomous positioning and navigation in the GPS-free environment in the flue. The obstacle avoidance system adopts a composite detection scheme of laser radar and infrared sensors and can realize response to obstacles such as protruding components and accumulated dust in the flue and can still maintain stable flight in the high-temperature and dusty environment.

[0024] Further, the imaging unit is installed on the unmanned aerial vehicle unit and comprises a high-definition visible light camera and an infrared camera. The visible light camera is used for shooting common photos of the water-cooled wall of the flue and assisting in judging the overall environment and structure in the flue. The infrared camera is used for capturing infrared thermal images of the water-cooled wall of the flue and acquiring surface temperature distribution information thereof to provide key data for damage detection.

[0025] Further, the control unit comprises a ground controller and an unmanned aerial vehicle on-board controller for realizing remote control of the unmanned aerial vehicle unit and the imaging unit. The ground controller can be connected with the unmanned aerial vehicle on-board controller through high-temperature-resistant optical fibers, and an operator can set flight parameters, flight routes and imaging parameters of the unmanned aerial vehicle on the ground, and monitor the flight state and imaging condition of the unmanned aerial vehicle in real time.

[0026] Further, the data processing unit comprises an image preprocessing module, a three-dimensional modeling module, a temperature analysis module and a contrast analysis module. The image preprocessing module is used for processing the received ordinary photos and infrared thermal images, including denoising, enhancement, correction and the like, to improve the image quality. The three-dimensional modeling module constructs a three-dimensional model of the flue water wall according to the flight trajectory of the unmanned aerial vehicle and the image data, to realize accurate positioning of the damage position. The temperature analysis module extracts and analyzes the temperature of the infrared thermal image, to identify a first area with a local temperature less than a first preset temperature threshold, which can be a crack or a position of wear thinning. The contrast analysis module compares the current detected infrared thermal image data with historical data stored in the database unit, to find a second area with a local temperature drop, to judge whether it is a potential damage position. Further, the database unit is used for storing infrared thermal image data, three-dimensional model data and related detection parameters and results of the flue water wall in previous detections. The database unit adopts a distributed storage architecture, has large capacity, high reliability and easy scalability, and can meet the data storage needs of long-term dynamic detection. At the same time, the database unit is also provided with a data management module, to realize classification, retrieval, updating and backup of data, to facilitate user query and use of historical data. Further, the optical fiber adopts a cladding structure of metal or ceramic and the like, and the cladding layer has a temperature resistance not less than 200℃ and a length adapted to the space of more than 30m of the flue.

[0027] Further, the optical fiber is connected to the tail of the unmanned aerial vehicle through an automatic take-up and pay-off reel, and is automatically released or recovered during flight, to avoid entanglement or jamming of the tow cable.

[0028] Further, the obstacle avoidance module comprises a combination of a laser radar and an infrared sensor, which can detect protruding components and dust accumulation in the flue, to realize dynamic path adjustment.

[0029] Further, in actual application, for the case that there is a large amount of smoke dust in the flue, a small smoke dust removal device such as a miniature fan can be installed on the unmanned aerial vehicle unit, to simply clean the surface of the water wall before shooting, to reduce the influence of smoke dust on the imaging quality. At the same time, a smoke dust recognition and removal algorithm is added in the image preprocessing module, to further improve the quality of the infrared thermal image. Further, the system can also interact with the production management system of the converter to associate the detected damage information with the production parameters (such as smelting time, molten steel temperature, cooling water flow, etc.) of the converter, analyze the relationship between damage generation and production parameters, and provide data support for optimizing the production process of the converter and reducing water-cooled wall damage. As shown in Figure 2 The dynamic detection method for the flue water-cooled wall damage of the converter waste heat boiler provided by the embodiment of the present application based on the above system comprises the following steps: Detection preparation stage: plan the flight route of the unmanned aerial vehicle according to the flue structure of the converter 8 and set the imaging parameters; Under the condition that the converter stops blowing oxygen, the flue temperature drops to about 200℃, the flue induced draft fan stops running, and the water-cooled wall cooling water circulation is maintained, the unmanned aerial vehicle is controlled to fly into the flue from the set take-off point along the preset flight route, avoid obstacles, and perform image acquisition; The imaging unit acquires the visible light image and infrared thermal image of the flue water-cooled wall, and the images are stored in the local storage of the unmanned aerial vehicle in real time and transmitted to the ground control unit via high-temperature resistant optical fibers; The image data is transmitted to the data processing unit to perform image preprocessing, three-dimensional modeling, temperature extraction analysis and pseudo-color processing; The temperature image of the current detection is compared with the historical image in the database by difference, the local temperature abnormal area is identified, and the potential damage is marked; Output the detection report and update the database for guiding subsequent maintenance or trend analysis.

[0030] Further, during the flight of the unmanned aerial vehicle, real-time communication and data transmission are realized through high-temperature resistant optical fibers connected to the ground controller, and the optical fibers adopt a metal or ceramic material cladding structure to adapt to the flue environment above 200℃.

[0031] Further, the identification of the temperature abnormal area is based on a second preset temperature threshold of 5℃ or above, and the damage point is determined after three-dimensional registration of the structure.

[0032] Further, the image processing includes median filtering, contrast enhancement and wavelet denoising, and the modeling adopts a Structure from Motion (SfM) algorithm to generate a three-dimensional model of the flue water-cooled wall.

[0033] Further, in the detection preparation stage, the flight route 10 of the unmanned aerial vehicle is planned and designed in advance according to the converter workshop design drawings and the actual situation on site, the flight route is planned according to the structural characteristics and size of the flue to ensure that the detection area of the entire flue water-cooled wall can be covered and blind areas are avoided, the planned route is stored in the unmanned aerial vehicle, and a fixed take-off point 7 is calibrated on the converter workshop platform.

[0034] According to the field situation, set the shooting parameters of the imaging unit, such as shooting interval, exposure time and other parameters, and store them in the imaging unit.

[0035] The data processing unit and the database unit are installed with relevant image processing and data processing software.

[0036] Further, the unmanned aerial vehicle take-off and flight detection includes: After the converter stops blowing oxygen and completes tapping, it is confirmed that the flue is in a hot state and there is cooling water circulating in the water-cooled wall pipe. Keep the flue induced draft fan running, wait for about 10 minutes, and the temperature in the flue will drop to about 200℃. The flue induced draft fan stops running to avoid airflow in the flue interfering with the flight of the unmanned aerial vehicle. The operator starts the unmanned aerial vehicle unit through the control unit, and the unmanned aerial vehicle flies into the converter waste heat boiler flue 9 from the preset take-off point according to the preset flight route.

[0037] The unmanned aerial vehicle will fly and shoot according to the planned route and transmit real-time data pictures, or the unmanned aerial vehicle can be manually controlled, and the shooting parameters can be manually set in real time. Further, the data processing and analysis stage includes: The photos stored in the imaging unit are transmitted to the data processing unit. The image preprocessing module processes the received ordinary photos and infrared thermal images: the ordinary photos are denoised and enhanced to exclude part of the image interference caused by smoke dust; the infrared thermal images are denoised, temperature calibrated and pseudo-color processed to improve the clarity and readability of the images. The three-dimensional modeling module uses the motion recovery structure (SfM) algorithm to construct a three-dimensional model of the flue water-cooled wall according to the flight trajectory data of the unmanned aerial vehicle and the processed image data, maps the temperature information of the infrared thermal image onto the three-dimensional model, and realizes the spatial positioning of the damage position.

[0038] The temperature analysis module performs temperature analysis on the processed infrared thermal images, extracts the temperature distribution data of the water-cooled wall surface, and identifies the first area with locally low temperature. According to the principle of heat conduction, when the water-cooled wall has cracks or is worn thin, its local heat dissipation performance changes, resulting in a lower temperature, so these areas are marked as potential damage areas. Further, the database comparison and damage confirmation stage includes: The contrast analysis module compares the current detected infrared thermal image temperature data and potential damage area information with historical data stored in the database unit. The database unit stores temperature distribution data and a three-dimensional model of previous detections. The contrast analysis module finds the second area with more local temperature drop compared with previous data through image registration and temperature difference calculation. Combining the first area with low local temperature identified by the temperature analysis module and the second area with temperature drop found by the contrast analysis module, the possible damage position that needs to be focused on is determined comprehensively, and is marked on the three-dimensional model. Further, the detection result output and maintenance arrangement stage includes: The data processing unit can view and export the detection results on the computer, including the position of the potential damage area, temperature information, three-dimensional model marking, and comparison analysis report with historical data. According to the detection results, the maintenance personnel can check and repair the marked potential damage position during the regular maintenance of the converter, and eliminate the safety hazards in time. Further, the database updating stage includes: After each detection is completed, the data processing unit stores the image data, temperature data, three-dimensional model and detection results of this detection to the database unit, and updates the database. With the increase of the number of detections, the database is continuously improved, providing more historical data for subsequent comparison analysis, and improving the accuracy and reliability of damage detection. The following is a specific operation example provided by an embodiment of the present application.

[0039] (1) System composition and parameter setting In this embodiment, the specific parameters of each component of the intelligent detection system for the flue of the converter waste heat boiler are as follows: 1. UAV unit: a multi-rotor unmanned aerial vehicle is used, the body is made of high-temperature resistant alloy material, which can withstand a high-temperature environment above 200℃. The high-precision inertial navigation system is equipped, the obstacle avoidance sensor uses laser radar and infrared, the detection distance is 0.5-10m, and obstacles with a diameter greater than 5mm can be detected. The maximum flight speed of the unmanned aerial vehicle is 5m / s, the endurance time is 40 minutes, which can meet the demand of one-time detection of the flue. 2. Imaging unit: all cameras are high-temperature resistant design, the resolution of high-definition visible light camera is 20 million pixels, the lens focal length is 10-30mm, and automatic focusing can be realized; the resolution of infrared thermal imager is 640x512, and the temperature measurement range is -20-300℃. 3. Control unit: The ground controller uses an industrial computer with a 15-inch high-definition display, Windows 10 operating system, and dedicated drone control and image processing software. The onboard controller uses a high-performance embedded processor with a 1GHz operating speed, allowing for fast response to control commands. The ground controller and onboard controller are connected by high-temperature-resistant optical fiber, with a temperature resistance of over 200°C.

[0040] 4. Data processing unit: A high-performance server with an Intel Xeon E5 processor, 128GB of memory, and 5TB of hard disk capacity is used. Image preprocessing software uses the OpenCV open source library for secondary development, implementing image denoising and enhancement functions. Three-dimensional modeling uses professional three-dimensional modeling software combined with self-developed algorithms for rapid modeling. Temperature analysis and comparative analysis modules are developed using Python programming language and machine learning algorithms to improve temperature analysis accuracy. 5. Database unit: MySQL database is used, supporting distributed storage and enabling automatic data backup and recovery. Database management software has user permission management, data query, and statistical analysis functions, making it easy for users to manage and use historical data. (2) Detection process Taking the detection of a 300t converter waste heat boiler flue water wall in a certain steel enterprise as an example, the detection process of the invention is described in detail: Step one: According to the design drawings of the converter waste heat boiler flue and the actual situation on site, the flight route of the unmanned aerial vehicle is planned on the ground controller.

[0041] Step two: After the converter completes a blow oxygen steelmaking and tapping (during which the induced draft fan is ventilated and the circulating cooling water cools the flue), the converter body is deflected to the tapping angle. At this time, the flue still maintains a certain temperature, and the water wall tube has normal circulation of cooling water. Turn on the ground controller and the onboard controller, and perform system self-checking to ensure that each unit is normal. The starting point is a fixed takeoff point planned in advance, the flight route covers all areas of the flue water wall, the distance between adjacent shooting points is 0.5m, and the images captured can be seamlessly spliced. The flight height of the unmanned aerial vehicle is set to 1-2m from the surface of the water wall, the flight speed is 2m / s, the infrared thermal imager shooting interval is 0.1s, and the high-definition visible light camera shooting interval is 1s. Step three: After 10 minutes of converter tapping, the flue induced fan is turned off, the cooling water circulation is maintained, the operator issues a take-off command from the ground controller, and the UAV takes off from the converter operating platform and flies into the flue according to the preset flight route. During the flight, the on-board controller automatically adjusts the flight attitude according to the deviation compared with the preset route to ensure that the UAV flies along the preset route. The obstacle avoidance sensor scans the front environment in real time, and when it detects protrusions or other obstacles in the flue, the UAV automatically slows down and adjusts the route to bypass the obstacles, and then continues to fly along the original route after avoiding the obstacles.

[0042] The imaging unit starts working according to the set parameters, and the high-definition visible light camera and the infrared thermal imager synchronously take pictures to obtain ordinary photos and infrared thermal images of the flue water-cooled wall, and transmit them to the ground controller through high-temperature-resistant optical fibers.

[0043] Step four: data processing and analysis: after receiving the data, the image preprocessing module immediately processes the images. For ordinary photos, the median filter algorithm is used to remove noise caused by smoke dust, and the contrast enhancement algorithm is used to improve the clarity of the images. For infrared thermal images, the wavelet transform denoising algorithm is used to remove noise interference, the temperature is calibrated according to the calibration parameters of the infrared thermal imager, the gray value of the image is converted into the actual temperature value, and the pseudo-color processing is performed to represent different temperatures with different colors, so that the temperature distribution is more intuitive.

[0044] The three-dimensional modeling module calculates the camera attitude of each image using the motion recovery structure algorithm according to the flight trajectory data (including position, attitude, time, etc.) of the UAV and the processed infrared thermal images and ordinary photos, and then generates point cloud data through dense matching to construct a three-dimensional grid model of the flue water-cooled wall. The temperature information of the infrared thermal image is mapped to the corresponding position of the three-dimensional grid model to form a three-dimensional model with temperature information.

[0045] The temperature analysis module performs point-by-point temperature analysis on the three-dimensional model with temperature information to find local low-temperature positions, i.e., the first area with a temperature less than the first preset temperature threshold. These areas are marked as potential damage areas, and their position coordinates and temperature values are recorded.

[0046] Step five: database comparison and damage confirmation: the comparison and analysis module retrieves the latest detection data of the flue water-cooled wall of the converter waste heat boiler from the database unit, including the temperature distribution data of the infrared thermal image and the three-dimensional model. The three-dimensional model of this detection is registered with the historical three-dimensional model to ensure that their spatial positions are consistent. Through the image registration algorithm, the infrared thermal image of this detection is accurately aligned with the historical infrared thermal image, and the temperature difference of the corresponding position is calculated.

[0047] The second preset temperature threshold is set to 5°C, and when the temperature of a certain area in this time is decreased by more than 5°C compared with the historical temperature, the area is marked as a second area of abnormal temperature decrease. The comparative analysis module superimposes and analyzes the first area of potential damage marked by the temperature analysis module and the second area of abnormal temperature decrease, and takes the intersection of the two as the possible damage position for focus. For the position appearing only in the first area of potential damage or the second area of abnormal temperature decrease, the structure characteristics and operation of the flue are comprehensively judged to exclude the temperature change caused by the cooling water flow fluctuation or local structure difference, and finally the position needing maintenance is determined.

[0048] Step six: detection result output and maintenance arrangement: the data processing unit generates a detection report, including the three-dimensional model of the flue water-cooled wall (with damage position marking), the temperature distribution chart of the potential damage area, the comparative analysis chart with historical data, and the detailed coordinates and temperature information of the damage position. The detection report is sent to the ground controller through the data transmission unit, and the operator can view, print or export the detection report on the ground controller.

[0049] The detection report is submitted to the equipment management department, which arranges the regular maintenance plan of the converter according to the detection report. During maintenance, the maintenance personnel check the flue water-cooled wall according to the damage position marked in the report, and use ultrasonic thickness gauge and penetration detection method to confirm the damage condition. For the positions confirmed to have cracks or wear thinning, timely repair welding is carried out to ensure the safe operation of the water-cooled wall.

[0050] Step seven: database update: after the completion of this detection, the data processing unit stores all the data of this detection, including the original image data, the processed image data, the three-dimensional model data, the temperature analysis result, the comparative analysis result and the detection report, in the database unit according to the time stamp and the flue number. The database management system automatically classifies and organizes the data, establishes index, and facilitates subsequent query and comparison. At the same time, the system backs up the database to prevent data loss.

[0051] In the subsequent converter production process, detection is carried out every other month according to the same process. With the increase of detection times, the historical data in the database is constantly enriched, and the accuracy and reliability of comparative analysis are further improved, which can more timely and accurately find the damage change of the flue water-cooled wall.

[0052] Further, in terms of temperature analysis algorithm, a machine learning model can be introduced, which is trained on a large amount of historical detection data and actual damage conditions, so that the model can more accurately identify the temperature characteristics corresponding to different types of damage (such as cracks, wear thinning), and improve the accuracy of damage detection. For example, for cracks, the temperature distribution around them may present a specific gradient change; for wear thinning areas, the temperature may present a relatively uniform low state, and the machine learning model can learn these features to achieve more accurate classification and identification.

[0053] The intelligent converter waste heat boiler flue detection method and system provided by the embodiments of the present application are described in detail above, and the principles and implementation manners of the present application are described by applying specific examples. The above embodiment description is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed, and the above description should not be understood as a limitation of the present application.

Claims

1. An intelligent detection system for converter waste heat boiler flue, characterized in that: include: UAV unit, imaging unit, control unit, data processing unit and database unit; The drone unit is used to achieve autonomous positioning and navigation in a GPS-free environment within the flue; The imaging unit is installed on the UAV unit and includes a visible light camera and an infrared camera. The visible light camera is used to take ordinary photos of the flue water-cooled wall, and the infrared camera is used to capture infrared thermal images of the flue water-cooled wall. The control unit is used to remotely control the drone unit and the imaging unit; The data processing unit is used to construct a three-dimensional model of the flue water-cooled wall based on the flight trajectory of the drone, ordinary pictures and infrared thermal images, identify abnormal areas by analyzing the temperature of the infrared thermal images, and search for corresponding historical data stored in the database unit through the infrared thermal images and the abnormal areas, so as to determine whether the abnormal areas are potential damage locations and mark them in the three-dimensional model.

2. The system according to claim 1, wherein: The data processing unit includes an image preprocessing module, a three-dimensional modeling module, a temperature analysis module and a comparative analysis module; The image preprocessing module is used to process ordinary photos and infrared thermal images; The three-dimensional modeling module is used to construct a three-dimensional model of the flue water wall based on the flight trajectory of the UAV, the processed ordinary photos and the infrared thermal images, and map the temperature information of the processed infrared thermal images to the corresponding positions of the three-dimensional model to form a three-dimensional model with temperature information; The temperature analysis module is used to perform point-by-point temperature analysis on the three-dimensional model with temperature information, find a first area where the local temperature is less than a first preset temperature threshold, and mark the identified first area as a potential damage area; The comparison and analysis module is used to compare the currently detected infrared thermal image data with the historical infrared thermal image data stored in the database unit, find the second area where the local temperature drops in the historical infrared thermal image, determine the possible damage location based on the first area and the second area, and make a comprehensive judgment based on the structural characteristics and operating conditions of the flue to finally determine the location that needs maintenance.

3. The system according to claim 2, characterized in that The comparative analysis module is used to retrieve the most recent inspection data of the water-cooled wall of the converter waste heat boiler flue from the database unit, including the temperature distribution data of the historical infrared thermal image and the historical three-dimensional model, and align the three-dimensional model of this inspection with the historical three-dimensional model to ensure that the spatial positions of the two are consistent, accurately align the infrared thermal image of this inspection with the historical infrared thermal image, and calculate the temperature difference at the corresponding position. When the current temperature of a certain area drops by more than a second preset temperature threshold compared with the historical temperature, the area is marked as the second area with abnormal temperature drop, the first area and the second area are superimposed and analyzed, and the intersection of the first area and the second area is taken as the possible damage location. For the location that only appears in the first area or the second area, a comprehensive judgment is made based on the structural characteristics and operation conditions of the flue to eliminate the temperature changes caused by fluctuations in cooling water flow or local structural differences, and finally determine the location that needs maintenance.

4. The system according to claim 3, characterized in that The control unit includes a ground controller and an onboard controller of the UAV. The ground controller is connected to the onboard controller of the UAV via a high-temperature resistant optical fiber, so that the operator can control and set the flight parameters, flight route and imaging parameters of the UAV on the ground through the ground controller, and monitor the flight status and imaging of the UAV in real time.

5. The system according to claim 4, characterized in that The data processing unit is also used to generate an inspection report, including a three-dimensional model of the flue water-cooled wall with damage location marks, a temperature distribution chart of the first area, a comparative analysis chart with historical data, and detailed coordinates and temperature information of possible damage locations.

6. A converter waste heat boiler flue intelligent detection method, characterized in that: include: Through remote control, the drone unit can achieve autonomous positioning and navigation in the flue without GPS, and take ordinary photos and infrared thermal images of the flue water-cooled wall; A three-dimensional model of the flue water-cooled wall is constructed based on the flight trajectory of the drone, ordinary pictures and infrared thermal images. Abnormal areas are identified by analyzing the temperature of the infrared thermal images. Corresponding historical data is found through the infrared thermal images and the historical data stored in the database unit of the abnormal areas to determine whether the abnormal areas are potential damage locations and mark them in the three-dimensional model.

7. The method according to claim 6, characterized in that The method of constructing a three-dimensional model of the flue water wall based on the flight trajectory of the drone, ordinary pictures, and infrared thermal images, identifying abnormal areas by analyzing the temperature of the infrared thermal images, and searching for corresponding historical data stored in the database unit using the infrared thermal images and the abnormal areas to determine whether the abnormal areas are potential damage locations and marking them in the three-dimensional model includes: Processing ordinary photos and infrared thermal images, constructing a 3D model of the flue water wall based on the UAV's flight trajectory, the processed ordinary photos, and the infrared thermal images. The temperature information of the processed infrared thermal images is then mapped to the corresponding positions of the 3D model, forming a 3D model with temperature information. Performing point-by-point temperature analysis on the three-dimensional model with temperature information, finding a first area where the local temperature is less than a first preset temperature threshold, and marking the identified first area as a potential damage area; The currently detected infrared thermal image data is compared with the historical infrared thermal image data stored in the database unit to find the second area where the local temperature drops in the historical infrared thermal image. The possible damage location is determined by the first area and the second area, and a comprehensive judgment is made based on the structural characteristics and operating conditions of the flue to finally determine the location that needs maintenance.

8. The method according to claim 7, characterized in that The currently detected infrared thermal image data is compared with the historical infrared thermal image data stored in the database unit to find the second area where the local temperature drops in the historical infrared thermal image, determine the possible damage location based on the first area and the second area, and make a comprehensive judgment based on the structural characteristics and operation status of the flue to finally determine the location that needs to be repaired, including: The most recent inspection data of the water-cooled wall of the converter waste heat boiler flue is retrieved from the database unit, including the temperature distribution data of the historical infrared thermal image and the historical three-dimensional model. The three-dimensional model of this inspection is aligned with the historical three-dimensional model to ensure that the spatial positions of the two are consistent. The infrared thermal image of this inspection is accurately aligned with the historical infrared thermal image, and the temperature difference of the corresponding position is calculated. When the current temperature of a certain area drops by more than the second preset temperature threshold compared with the historical temperature, the area is marked as the second area with abnormal temperature drop. The first area and the second area are superimposed and analyzed, and the intersection of the first area and the second area is taken as the possible damage location. For the location that only appears in the first area or the second area, a comprehensive judgment is made based on the structural characteristics and operation conditions of the flue to eliminate the temperature changes caused by fluctuations in cooling water flow or local structural differences, and finally the location that needs maintenance is determined.

9. The method according to claim 8, characterized in that The method further comprises: Generate an inspection report, including a 3D model of the flue water-cooled wall with damage locations marked, a temperature distribution chart for the first area, a comparison analysis chart with historical data, and detailed coordinates and temperature information of possible damage locations.

10. The method according to claim 9, characterized in that The method further comprises: The machine learning model is trained on historical inspection data and actual damage conditions to identify the temperature characteristics corresponding to different types of damage. The trained machine learning model is then used to identify the three-dimensional model with temperature information to obtain potential damage areas.

Images