A ladle state monitoring system and method of multi-source sensing fusion in high temperature environment

Abstract

This invention provides a ladle condition monitoring system and method based on multi-source sensor fusion under high-temperature conditions. It employs multiple sensing methods, including laser 3D scanning, infrared thermal imaging, and visible light vision, to comprehensively monitor the condition of the refractory lining of the ladle. The system includes a multi-sensor combination probe mounted on a robotic arm, along with its high-temperature protective device, and a data processing and control module. This module acquires the ladle's geometric morphology, temperature field distribution, and image information under conditions of high temperature, red-hot ladle heat, and smoke interference. It then uses a data fusion algorithm to determine state parameters such as ladle lining thickness, abnormal temperature areas, and slag adhesion. The system can detect weak areas or overheating signs in the ladle lining in real time and provide early warning signals, preventing accidents such as ladle leakage caused by excessive lining erosion or temperature runaway. Cross-validation of multi-source information improves the accuracy and robustness of the monitoring results, achieving intelligent and digital management of the ladle condition.

Description

Technical Field

[0001] This invention relates to the field of steelmaking equipment condition monitoring technology, and in particular to a ladle condition monitoring system and method based on multi-source sensor fusion under high-temperature conditions, used to acquire the integrity status and operational safety parameters of the ladle lining in real time. Background Technology

[0002] As a container for transferring molten steel, the safe operation of the ladle is crucial to steel production. During use, the high-temperature molten steel corrodes and thermally shocks the refractory lining of the ladle. If excessive wear or localized overheating of the lining goes undetected, it can lead to serious accidents such as molten steel leakage (ladle leak). Therefore, online monitoring of the ladle's condition has always been a key challenge and focus of the metallurgical industry.

[0003] Traditional ladle monitoring methods primarily rely on manual inspection and single-sensor detection. For example, after a ladle is taken out of service, workers visually inspect the lining thickness. This method is not only subjective and inaccurate, but also requires personnel to work near high-temperature environments, posing safety hazards. Some steel mills have attempted to use fixed infrared thermometers to monitor the ladle's outer wall, hoping to detect abnormal temperatures during the molten steel casting process to infer weak areas in the lining. However, relying solely on outer wall temperature monitoring cannot accurately pinpoint the extent and location of lining erosion, and is susceptible to interference from ambient flames and steam, leading to false alarms. Furthermore, some companies use laser rangefinders or 3D scanners to measure the shape of the cooled ladle lining, but this requires the ladle to be cooled to a lower temperature before this can be done, making real-time monitoring of each ladle turnover impossible.

[0004] With the development of sensing technology, multi-source information fusion monitoring has become an important trend in improving detection reliability. In the steel ladle field, combining data from multiple sensors can provide a more comprehensive understanding of the ladle's condition: for example, 3D laser scanning can directly obtain the thickness distribution of the refractory lining, infrared thermal imaging can reveal temperature hotspots on the lining surface, and visual images can help identify ladle numbers and lining surface anomalies. However, achieving stable measurement and data fusion from multiple sensors in the high-temperature, high-radiation, and dusty steelmaking environment still faces challenges: sensors need to operate reliably in environments approaching hundreds or even thousands of degrees Celsius, requiring effective cooling and insulation measures; the data formats and spatiotemporal resolutions acquired by different types of sensors vary, and how to align and fuse them to make accurate judgments is also a technical difficulty.

[0005] Therefore, it is necessary to study a ladle condition monitoring system and method based on multi-source sensor fusion under high-temperature conditions to address the shortcomings of existing technologies, solve or mitigate one or more of the above-mentioned problems, and achieve automatic, accurate, and real-time assessment of the ladle condition. Summary of the Invention

[0006] In view of this, the present invention provides a ladle condition monitoring system and method based on multi-source sensor fusion under high temperature environment. By integrating multiple sensors and combining them with intelligent data fusion algorithms, it achieves comprehensive monitoring of the ladle lining condition and operating parameters, thereby overcoming the shortcomings of existing technologies such as limited coverage and susceptibility to interference caused by single detection methods.

[0007] On one hand, the present invention provides a ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions, the ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions comprising: The system includes a robotic arm, multi-source sensor components, cooling and protection devices, a data processing and control module, and a display and alarm unit. The robotic arm device is used to position the multi-source sensor assembly to a detection location inside a steel ladle that is in a high-temperature, red-hot state; The multi-source sensor assembly is installed at the end of the robotic arm device and includes a first imaging sensor, a second imaging sensor and a third imaging sensor, which are used to collect point cloud geometric data, surface temperature distribution data and visible light image data of the steel ladle refractory lining, respectively. The cooling and protection device covers the multi-source sensor assembly and includes an internal water-cooling circuit, a high-temperature insulation layer and an external gas purging structure, which is used to maintain the operating temperature of each sensor and keep the optical window clean in a high-temperature and dusty environment. The data processing and control module is connected to the multi-source sensor component and is used to perform spatiotemporal alignment and fusion processing on the three-source data to generate a three-dimensional point cloud model of the lining with temperature attributes, and output the ladle lining status information according to the preset judgment rules. The display and alarm unit is used to present the status information of the ladle lining and to issue an alarm signal when an abnormality is detected. In addition to the aspects and any possible implementations described above, a further implementation is provided in which the first imaging sensor is a three-dimensional laser sensor for acquiring point cloud geometric data of the refractory lining of the steel ladle; The second imaging sensor is an infrared thermal imaging sensor, used to collect surface temperature distribution data of the refractory lining of the steel ladle; The third imaging sensor is a visible light camera, used to collect visible light image data of the refractory lining of the steel ladle; The data processing and control module includes a multi-source data fusion unit, which is used for: The point cloud data of the three-dimensional laser sensor is used as the geometric reference coordinate system; The temperature pixel matrix of the infrared thermal imaging sensor is mapped to the point cloud coordinate system through pre-calibrated intrinsic and extrinsic parameters to form a temperature attribute point cloud. The image pixels of the visible light camera are mapped to the point cloud coordinate system through pre-calibrated intrinsic and extrinsic parameters to form an image attribute point cloud; In addition to the aspects and any possible implementations described above, a further implementation is provided in which the internal water-cooling circuit is coiled around a cooling water pipe surrounding the multi-source sensor assembly; The high-temperature insulation layer is wrapped between the housing of the multi-source sensor assembly and the water-cooling circuit; The external gas purging structure includes an annular nozzle disposed at the front end of the multi-source sensor assembly, used to spray compressed air into the sensor's field of view to form an air curtain. In addition to the aspects described above and any possible implementations, a further implementation is provided in which the data processing and control module further includes a state determination unit, the state determination unit being used for: The residual thickness of each part of the steel ladle lining is calculated based on the temperature attribute points. Identify abnormal surface temperature regions based on the temperature attribute point cloud; The slag adhesion area is identified based on the image attribute point cloud.

[0008] As described above and in any possible implementation, a further implementation is provided, wherein the state determination unit includes an interference filtering subunit, the interference filtering subunit being used for: When the highest temperature zone is located in the ladle top edge opening area or the tapping area, the temperature anomaly is determined to be due to molten steel residue interference; the temperature influence of this zone is ignored in the thickness assessment. Based on the brightness characteristics of the image attribute point cloud, when there are bright residual molten steel image features in the top edge opening area or the steel outlet area, it is further confirmed that the high temperature is due to residual molten steel interference.

[0009] In addition to the aspects described above and any possible implementation, a further implementation is provided in which the state determination unit is also used to: determine that there is a serious risk of corrosion in a certain area when the residual thickness of a certain area is lower than a first threshold and the surface temperature of the area is higher than a second threshold; generate a corresponding abnormality identifier and send it to the display and alarm unit.

[0010] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the data processing and control module further includes a ladle identity management unit, which is used to: perform optical character recognition based on the ladle number image captured by the visible light camera; obtain ladle identification information through an interface with the manufacturing execution system; and associate and store the current monitoring data with the unique identifier of the corresponding ladle.

[0011] In accordance with the aspects and any possible implementations described above, a further implementation is provided in which the display and alarm unit includes an industrial display screen and an audible and visual alarm, the industrial display screen being used for: Real-time display of a 3D point cloud model or unfolded cross-sectional view of the ladle lining with temperature pseudo-color overlay; Highlight and mark abnormal areas; Displays the ladle number, cumulative number of uses, current thinnest point thickness and location, and highest temperature point and location.

[0012] As described above and in any possible implementation, a method for monitoring the condition of a ladle using multi-source sensor fusion under high-temperature conditions is further provided. Monitoring is performed using a multi-source sensor fusion ladle condition monitoring system under high-temperature conditions. The method includes the following steps: Step 1: Steel ladle identification, obtaining the identity information of the steel ladle to be inspected; capturing the number image of the steel ladle through a visible light camera and identifying the steel ladle number, or retrieving the steel ladle number and historical usage data from the factory database; Step 2: Laser scanning. Drive the robotic arm to extend the 3D laser sensor probe into the ladle and perform a 360° full-coverage scan of the refractory lining surface of the ladle to collect high-density point cloud data to obtain the geometry of the lining. Step 3: Temperature measurement. Use an infrared thermal imaging sensor to obtain the temperature distribution map of the steel ladle lining surface, and obtain accurate and reliable temperature data through appropriate filtering and calibration. Step 4: Image acquisition. Use a visible light camera to acquire visible light images of the ladle lining. Under specific conditions, image analysis can be used to identify color changes or slag adhesion on the lining surface. Step 5: Data fusion. The data processing module aligns and fuses the data from the above-mentioned sensor sources in time and space, associates the temperature data with the three-dimensional geometric model, and generates a comprehensive model that includes both lining thickness information and temperature information. This model is then combined with visible light images to corroborate suspicious areas. Step 6: Status determination. Analyze the fused model data and evaluate the status of the ladle lining, including calculating the residual thickness of refractory material in each part, detecting hot spots with excessive temperature, and identifying slag accumulation at the ladle opening or inner wall; determine whether there are any abnormalities based on the preset safety threshold, including the thickness of the lining at a certain point being lower than the threshold or the temperature rising abnormally. Step 7: Output the results. The analyzed ladle status information is output through the display and alarm unit. The display interface presents a three-dimensional / two-dimensional status diagram of the ladle lining, marking weak areas and abnormal temperature areas. When there is a safety hazard, an audible and visual alarm is triggered to prompt the operator to take appropriate measures.

[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step 5, a one-to-one correspondence is established between the temperature matrix collected by the infrared thermal imaging sensor and the three-dimensional laser point cloud, including calibrating the coordinate systems of the two types of sensors to ensure that each temperature pixel is mapped to the correct position on the three-dimensional point cloud model of the steel ladle lining, thereby obtaining a point cloud model with temperature attributes. In step 6, when the highest temperature area is detected at the top opening of the ladle or the bottom outlet of the ladle, it is determined that the high temperature may be caused by molten steel residue rather than thinning of the lining. Therefore, the abnormal temperature point is ignored in the lining corrosion assessment. When it is found that the residual thickness of a certain lining area is lower than the preset value and the temperature at the corresponding location is higher than the warning threshold, it is comprehensively judged that the refractory material in that area has been severely deteriorated, and a warning message that it needs to be repaired and replaced immediately is output.

[0014] Compared with the prior art, the present invention can achieve the following technical effects: 1. Multi-source sensor fusion overcomes the defect of single sensor being susceptible to interference and greatly improves detection accuracy. Laser thickness measurement and thermal imaging complement each other, quantitatively measuring the thickness change of the lining and qualitatively reflecting the heat load status. The combination of the two can more reliably identify safety hazards. 2. The innovative high-temperature resistant sensor protection technology enables detection to be carried out directly when the ladle is hot, without waiting for the ladle to cool down, thus realizing real-time online monitoring during the ladle turnover process and ensuring the safety of each heat of molten steel transportation; 3. Automatic ladle identification and data networking functions facilitate unified management of numerous ladles in the factory. It can record the historical corrosion curve and repair status of each ladle, assisting in optimizing refractory material replacement plans and reducing maintenance costs. 4. This invention provides an intelligent ladle monitoring solution for the steel smelting process, which can significantly improve production safety and management efficiency.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the technical effects described above at the same time. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the overall structure of a multi-source sensor fusion ladle condition monitoring system provided in one embodiment of the present invention (components include a robotic arm, sensor probes, main control cabinet, and operation terminal).

[0018] Figure 2 This is a schematic diagram of a sensor cooling protection device provided in one embodiment of the present invention, showing the protective design such as water cooling pipeline, heat insulation layer and air purging.

[0019] Figure 3 This is a flowchart of a ladle condition monitoring method provided in one embodiment of the present invention, which includes steps such as ladle identification, laser scanning, temperature measurement, image acquisition, data fusion, condition determination and result output.

[0020] Figure 4 This is a schematic diagram of the display interface of a monitoring system provided in one embodiment of the present invention. It presents the thickness and temperature distribution of the ladle lining in real time, and highlights abnormal areas and important parameters, such as remaining thickness or corrosion amount. Each grid area corresponds to a part of the ladle lining and is used to intuitively represent the wear condition.

[0021] Among them, 1-robot arm, 2-multi-source sensor probe, 3-data processing and control cabinet in the main control room, 4-operation terminal, 2a-360° lidar, 2b-infrared thermal imager, 2c-industrial camera. Detailed Implementation

[0022] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0024] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0025] This invention provides a ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions. The system structure, functional implementation of each component, and data processing flow are described in detail below with reference to specific embodiments.

[0026] During the continuous casting process in steelmaking, the steel ladle is exposed to a high temperature of around 1600℃ for extended periods. Under repeated erosion by molten steel, thermal shock, and chemical corrosion, its refractory lining gradually thins, developing localized corrosion pits, slag accumulation, and even partial spalling and detachment. If severely damaged areas are not detected and repaired or removed from the production line in time, ladle penetration accidents can easily occur, causing significant safety and economic losses. The system provided in this embodiment uses a robotic arm to safely deliver multi-source sensor components into the high-temperature, red-hot steel ladle. Under harsh conditions of strong radiation, pervasive smoke and dust, and extremely high temperatures, it simultaneously collects data on the lining's geometry, temperature field, and surface condition. Through data fusion and intelligent judgment, it outputs reliable information on the health status of the ladle lining.

[0027] The system mainly consists of five core parts: a robotic arm device, a multi-source sensor assembly, a cooling and protection device, a data processing and control module, and a display and alarm unit. These components work together to complete a closed loop from positioning and data acquisition to status determination.

[0028] As the motion execution component of the system, the robotic arm's main function is to accurately and safely deliver the sensor assembly mounted at its end effector into the designated detection position inside the ladle, and then quickly withdraw it from the high-temperature zone after the detection is completed. Typically, the robotic arm adopts a six-degree-of-freedom articulated industrial robot structure, with its base fixed to a dedicated platform or overhead crane beam above the ladle. The arm span design must cover the maximum opening diameter of the ladle and allow for a safety margin. In one embodiment, the robotic arm's end flange connects to the sensor assembly's mounting base via a quick-change interface, facilitating maintenance and switching between different sensor configurations. The robotic arm's motion trajectory is generated from a pre-planned offline programming path or online teaching path. Path planning must fully consider key obstacle avoidance points such as the upper edge of the ladle lining, the slag line area, and bottom corners, and incorporate multi-level safety limit and emergency stop mechanisms.

[0029] The multi-source sensor assembly is the core of the system information acquisition. It is installed at the front end of the robotic arm end effector and includes the first imaging sensor, the second imaging sensor and the third imaging sensor. The first imaging sensor, the second imaging sensor and the third imaging sensor are three sensing devices based on different principles. The first imaging sensor is a three-dimensional laser sensor used to collect point cloud geometric data of the refractory lining of the steel ladle; The second imaging sensor is an infrared thermal imaging sensor, used to collect surface temperature distribution data of the refractory lining of the steel ladle; The third imaging sensor is a visible light camera, used to collect visible light image data of the refractory lining of the steel ladle.

[0030] In one implementation, the three-dimensional laser sensor employs the time-of-flight or phase-based method, emitting near-infrared laser pulses onto the surface of the ladle lining and receiving diffuse reflection echoes. By calculating the time of flight or phase difference, the three-dimensional coordinates of each sampling point are obtained, forming high-density point cloud data. Its field of view typically covers 45° to 90°, and a single scan can acquire hundreds of thousands to millions of points. The infrared thermal imaging sensor operates in the long-wave infrared 8–14 μm band, enabling non-contact measurement of object surface temperature. It outputs a temperature matrix image with a temperature resolution on the order of 0.1°C and a spatial resolution of 640×480 or higher pixels. The visible light camera uses an industrial-grade color CMOS sensor with a resolution of no less than 1920×1080, equipped with automatic aperture and electronic shutter functions, to acquire optical information such as the true color texture, slag morphology, and crack orientation of the ladle lining surface.

[0031] Because the ladle is extremely hot and filled with high-temperature fumes and splashing steel slag particles, the optical windows of the three types of sensors are easily contaminated or their internal core components may overheat and fail due to thermal radiation. Therefore, a dedicated cooling and protection device must be configured. This cooling and protection device is a cylindrical or frustoconical shell that covers the main body of the multi-source sensor assembly, with optical windows reserved only in the field of view of each sensor.

[0032] In one embodiment, the cooling protection device includes a coiled water-cooling circuit. The water-cooling circuit uses a thin-walled stainless steel or copper alloy coil tightly wound around the outer surface of the sensor assembly housing. Cooling water is introduced from an external industrial circulating water system, boosted to 0.4–0.8 MPa by a booster pump, and then enters the coil, carrying away heat conducted from the sensor housing and internal components. The water then flows back to the cooling tower or heat exchange station via a return pipe. The coil design ensures a water flow velocity between 1.0 and 2.5 m / s to guarantee heat exchange efficiency while avoiding excessive water resistance and vibration. Quick-connect fittings and triple monitoring of temperature, pressure, and flow rate are installed at the cooling water inlet and outlet. If any parameter exceeds the safe range, an alarm is triggered, and the robotic arm insertion action is stopped.

[0033] A high-temperature insulation layer is further wrapped around the outside of the water-cooling circuit. This insulation layer preferably employs a multi-layered composite structure; for example, the innermost layer is a 5-10 mm thick microporous aluminosilicate fiber felt, the middle layer is a 15-25 mm thick aerogel felt, and the outermost layer is a 3-5 mm thick ceramic fiber cloth reinforced with stainless steel wire mesh. The function of the insulation layer is to block external radiative and convective heat (around 1600°C) as much as possible, maintaining the temperature inside the sensor assembly cavity below 80°C, thereby protecting the internal electronic components and optical elements for long-term stable operation.

[0034] At the very front of the cooling and protection device, around the optical windows of each sensor, an external gas purging structure is installed. This structure includes a ring of nozzles, typically with an inner diameter of 0.8–1.5 mm, evenly distributed around the circumference of the window. Compressed air, after being filtered to remove oil and water, is intermittently or continuously ejected at a pressure of 0.3–0.6 MPa, controlled by a solenoid valve, forming a high-speed air curtain in front of the sensor's field of view. This air curtain serves two purposes: firstly, it prevents dust particles from directly impacting the optical glass; secondly, it utilizes the Bernoulli effect to carry away dust near the window, thus maintaining the optical window's cleanliness for an extended period. In one possible implementation, the purging air path is also connected in parallel with a high-pressure nitrogen line as an emergency cleaning measure. When the effectiveness of compressed air deteriorates, it can be briefly switched to nitrogen for better purging results.

[0035] With the triple protection measures of internal water cooling, heat insulation layer barrier and external air curtain purging, the multi-source sensor assembly can work continuously and stably for more than 15 to 30 minutes in the high temperature and dust environment inside the ladle, meeting the time requirement for a complete ladle lining inspection.

[0036] The data processing and control module is the system's computing and decision-making hub, typically deployed in an industrial control cabinet or edge server. It interacts with multi-source sensor components in real time via gigabit Ethernet, fiber optics, or industrial camera interfaces. This module receives raw data streams from 3D laser sensors, infrared thermal imaging sensors, and visible light cameras. It first performs necessary preprocessing, then completes the spatiotemporal alignment and information fusion of the multi-source data, ultimately generating a 3D lining model with temperature and image attributes that can be used for state determination.

[0037] In one implementation, the workflow of the multi-source data fusion unit is as follows.

[0038] First, the point cloud data from the 3D laser sensor is used as the geometric reference coordinate system for the entire fusion process. Because the laser point cloud directly provides high-precision 3D geometric position information, its coordinate system is naturally suitable as a common reference frame. Before fusion, the point cloud data typically undergoes preprocessing such as downsampling, stereo filtering, and outlier removal to reduce subsequent computation and improve data quality.

[0039] Next, the temperature pixel matrix of the infrared thermal imaging sensor is mapped onto the aforementioned laser point cloud coordinate system to form a point cloud with temperature attributes. Specifically, the intrinsic and extrinsic parameters of the infrared thermal imaging sensor and the 3D laser sensor are pre-calibrated in the laboratory using a calibration plate or a dedicated calibration target to obtain the rotation matrix R, translation vector T, and possible distortion coefficients between them. Before each on-site inspection, a robotic arm can control the sensor assembly to perform a rapid online recalibration of the known geometric features of the ladle's upper edge to compensate for errors caused by thermal deformation and minor mechanical displacements. After calibration, for each effective temperature pixel in the infrared image, based on its pixel coordinates (u,v) and the projection relationship obtained from calibration, it is back-projected into 3D space to find the nearest neighbor point to the laser point cloud or to assign corresponding temperature values ​​to several nearest points through interpolation, thereby generating a point cloud with temperature attributes.

[0040] Similarly, the RGB images from the visible light camera are mapped to the same laser point cloud coordinate system using pre-calibrated intrinsic and extrinsic parameters. Since visible light images typically have higher spatial resolution, back projection combined with bilinear interpolation is often used during the mapping process to assign RGB color values ​​to the nearest laser point or its neighboring points, forming an image attribute point cloud. After this processing, the three data sources are precisely aligned geometrically, and the same physical location point simultaneously possesses three attributes: geometric coordinates, surface temperature, and visible light color texture, providing a rich information foundation for subsequent state determination.

[0041] It should be noted that because the surface of the ladle lining generates strong red heat radiation at high temperatures, the images captured by the visible light camera are often reddish overall and have low contrast. To improve the subsequent slag recognition effect, before mapping the RGB values ​​to the point cloud, gamma correction, histogram equalization, or color temperature correction based on a reference white field can be optionally performed on the original visible light image to partially restore the true color characteristics of the surface.

[0042] In one embodiment, when a large amount of floating dust inside the ladle causes local voids in the laser point cloud, the system can use edge information in the visible light image and continuous temperature gradient in the infrared image to perform geometric inference to complete the void area of ​​the point cloud, thereby improving the integrity of the overall three-dimensional model.

[0043] Through the above multi-source data fusion process, the system finally obtains a high-precision three-dimensional point cloud model of the ladle lining with temperature and image attributes. This model contains not only the real geometric shape, but also superimposed surface temperature distribution and visible light texture information, providing a unified data foundation for subsequent residual thickness assessment, temperature anomaly detection, and slag adhesion identification.

[0044] Based on the point cloud model with temperature and image attributes output by the aforementioned multi-source data fusion unit, the state determination unit in the data processing and control module continues to quantitatively assess and identify anomalies in the health status of the ladle lining.

[0045] The residual thickness of various parts of the ladle lining is calculated based on the temperature attribute point cloud. During the service life of the ladle refractory lining, the residual thickness is one of the most direct indicators for determining whether repair or scrapping is necessary. Since direct thickness measurement is difficult under high-temperature, red-hot conditions, this system uses the physical correlation between temperature field and thickness for indirect inference. The working layer of the ladle lining is typically composed of magnesia-carbon bricks or alumina-magnesia-carbon bricks, whose thermal conductivity varies with temperature and material degradation. When the lining thickness decreases, the heat from the molten steel is more easily transferred through the refractory layer, leading to a significant increase in the outer surface temperature of that area. Therefore, by analyzing the surface temperature values ​​of each region in the temperature attribute point cloud, and combining this with the current temperature of the molten steel in the ladle, pouring time, ladle wall thickness design value, and historical monitoring data, the residual thickness can be estimated.

[0046] Specifically, in one implementation, the temperature attribute point cloud is first divided into several typical regions according to the ladle geometry, such as the bottom region, slag line region, upper cylinder region, tapping port region, and top edge opening region. For each region, its temperature distribution statistical features are extracted, including average temperature, maximum temperature, temperature gradient, and the percentage of pixels with temperatures higher than a certain benchmark value. Then, based on a pre-established thickness-temperature correspondence table or empirical curve, the current temperature features are inverted to obtain the estimated residual thickness of the corresponding location. For example, when the average temperature of the slag line region is more than 80°C higher than normal mid-service temperature, and the maximum temperature point continues to appear, it can be determined that the residual thickness of this region is below the safety threshold and requires close monitoring.

[0047] In another embodiment, to improve the regional specificity of thickness estimation, the system can correlate and model the temperature attribute point cloud data from multiple historical monitoring of the ladle with the residual thickness data measured during actual ladle repair. By pairing and analyzing the temperature characteristics and corresponding thicknesses of the same ladle at different service cycles, the mapping relationship between regional temperature and thickness is gradually optimized, making the estimation results gradually approach the true value. This adaptive correction method based on the ladle's own historical data can effectively reduce the systematic errors caused by differences in ladle materials and masonry processes.

[0048] The temperature attribute point cloud is used to identify abnormal surface temperature regions. Surface temperature anomalies typically indicate defects such as severe localized corrosion of the lining, through-cracks, or embedded foreign objects. The identification process first sets multiple temperature threshold levels; for example, level one anomalies are those exceeding the normal average temperature by 50°C, level two anomalies by exceeding 80°C, and level three anomalies by localized hotspots exceeding 120°C or approaching the temperature of molten steel. Then, threshold filtering is applied to each point in the temperature attribute point cloud, marking point groups exceeding each threshold level. Connectivity analysis is then performed to cluster spatially continuous high-temperature points into anomalous regions.

[0049] In one possible implementation, the identification of abnormal areas also incorporates temperature gradient information for auxiliary judgment. If a local area experiences a drastic temperature change, i.e., the temperature difference between adjacent points exceeds 30℃ / cm, it is marked as a potential crack or pre-stripping area even if the average temperature of that area has not reached the highest threshold. This gradient-sensitive identification strategy helps to detect hidden defects in advance and avoids missed detections caused by relying solely on absolute temperature values.

[0050] The image attribute point cloud is used to identify slag adhesion areas. Slag adhesion is a common phenomenon in ladle linings, referring to the layer of steel slag that adheres to the surface of the refractory material after molten steel solidifies. An excessively thick slag layer can affect the purity of the molten steel poured in the next application and may detach during subsequent use, causing inclusion defects. The image attribute point cloud provides rich color, brightness, and texture information, which can be used to distinguish between normal refractory material surfaces and slag-coated surfaces.

[0051] Specifically, areas with slag adhesion typically exhibit a brighter metallic luster, more pronounced flow patterns, and strong localized reflections compared to normal brick surfaces. In one implementation, the RGB values ​​of the image attribute point cloud are first converted to HSV or LAB color space to enhance the distinction between brightness and hue. Then, brightness thresholds and red channel proportion thresholds are set to classify the points in the point cloud, marking points with brightness above a certain value and a dominant red component as suspected slag adhesion points. Further, by analyzing the spatial connectivity of these points and their texture contrast with surrounding points, isolated bright spots are filtered out, retaining only connected regions with an area greater than a certain threshold as slag adhesion areas.

[0052] Preferably, temperature attributes can be used for cross-validation during the slag identification process. Normal slag-covered areas, due to the steel slag layer, typically have a lower surface temperature than exposed refractory brick areas. Therefore, when a bright area corresponds to a lower temperature value, the confidence level in its slag-covered properties can be increased; conversely, if a bright area corresponds to an abnormally high temperature, it is more likely a false appearance caused by residual molten steel or localized molten steel penetration, requiring further interference filtering in subsequent stages.

[0053] When the highest temperature area is located in the ladle top edge opening area or the tapping area, the temperature anomaly is judged as interference from molten steel residue. After tapping, a small amount of molten steel or slag often remains near the top edge opening and the tapping area. These high-temperature residues generate strong localized high-temperature radiation, with temperatures approaching or exceeding 1600℃, the temperature of molten steel. If not distinguished, these high-temperature points are easily misjudged as lining perforation or severe corrosion, leading to false alarms or unnecessary emergency ladle repairs.

[0054] In one embodiment, the system first determines whether the highest temperature point or the highest temperature region falls within a preset top edge opening interference zone or a steel outlet interference zone based on the geometric coordinates of the point cloud. The range of these interference zones is predefined using a three-dimensional geometric model of the ladle, such as an annular region extending downwards from the top edge opening by 150mm to 300mm, and a fan-shaped region with a radius of 400mm to 600mm around the steel outlet. When the highest temperature point falls within these regions, it is automatically marked as a molten steel residue interference category, and the weight of this region in subsequent thickness assessment and corrosion risk determination is reduced.

[0055] The temperature influence of this region is ignored in the thickness assessment. To avoid a systematic overestimation of the overall residual thickness due to the high temperature of residual molten steel, the system removes regions identified as interfering with the thickness calculation domain during thickness inversion, or replaces their temperature values ​​with the average temperature value of adjacent non-interfering regions before including them in the calculation. This regional shielding ensures that the thickness estimation results primarily reflect the thermal conductivity of the refractory lining itself, rather than the interference effect of residual molten steel.

[0056] Based on the brightness characteristics of the image attribute point cloud, when bright molten steel residue image features are present in the top edge opening area or the tapping area, it is further confirmed that the high temperature is due to molten steel residue interference. Molten steel residue typically exhibits strong metallic reflection and a bright flow pattern in visible light images, which is clearly different from the rough, dark red surface of slag. In the interference filtering subunit, brightness histogram features, reflective spot density, and local contrast are extracted from the image attribute point cloud of the top edge and tapping areas. When these indicators exceed the preset molten steel residue judgment threshold, even if the temperature in that area is not completely excluded by geometric location, its high temperature can be ultimately confirmed as molten steel residue interference, thereby avoiding misjudgment as lining damage.

[0057] For example, during one monitoring session, a hot spot with a temperature as high as 1580℃ was found near the top edge opening of the ladle. From a temperature perspective alone, it would easily be identified as a risk of ladle penetration. However, based on geometric location, the hot spot was determined to be within 180mm below the top edge opening. At the same time, image attributes showed that there were multiple bright reflective spots and flowing metal textures in this area. The system ultimately determined it to be interference from molten steel residue, and did not trigger a severe corrosion alarm, thus avoiding unnecessary shutdown for inspection.

[0058] In another possible implementation, when the interference filtering subunit still cannot clearly determine the nature of a certain high-temperature area, the system can record all three-source characteristics of the area, including temperature value, geometric location, brightness characteristics and color distribution, and mark it as "high-temperature area to be confirmed" in the current monitoring report, and increase the sampling density in the next monitoring so as to further determine its true nature through time-series comparison.

[0059] When the residual thickness of a region is below a first threshold and the surface temperature of that region is above a second threshold, that region is considered to have a serious risk of corrosion. The first and second thresholds are typically set based on the ladle design specifications, the type of refractory material, and the specific safety strategy of the steel plant. For example, for magnesia-carbon lined ladles, the first threshold can be set to 30%–40% of the design thickness, and the second threshold can be set to the normal average temperature plus 100°C–150°C. When a region simultaneously meets both conditions—a residual thickness below the first threshold and a surface temperature above the second threshold—it indicates that the refractory layer in that area has been severely thinned, and heat transfer is abnormally intense, posing a high risk of ladle penetration.

[0060] In one implementation, the system further calculates the risk level for areas that meet the above conditions. For example, if the thickness is less than 80% of the first threshold and the temperature is more than 150°C above the second threshold, it is determined to be extremely high risk; if the thickness is less than the first threshold but the temperature is only 50°C above the second threshold, it is determined to be moderate risk. This tiered assessment allows on-site maintenance personnel to arrange different treatment measures based on the risk level, such as local repairs, overall thermal spraying, or early shutdown for major overhauls.

[0061] The system generates a corresponding anomaly identifier and sends it to the display and alarm unit. Once a severely erosion-risk area is identified, the status determination unit immediately generates an anomaly identifier data packet containing information such as location coordinates, risk level, estimated residual thickness, and maximum temperature, and pushes it to the display and alarm unit in real time via industrial Ethernet. Simultaneously, to facilitate traceability and analysis, the system also associates and stores the anomaly identifier with the complete 3D model and original data of this monitoring, forming an anomaly event archive available for later review.

[0062] The ladle identification management unit is used to ensure that each monitoring data can be accurately matched to a specific ladle, avoiding data mismatch or confusion.

[0063] Optical character recognition is performed based on images of the ladle serial numbers captured by the visible light camera. The outer wall of the ladle is typically painted or inlaid with a unique ladle serial number, such as "SP-0428" or "LF-156," which remains visible throughout the ladle's transport process. The visible light camera additionally captures multiple frames of images of the serial number area on the ladle's outer wall before or after the robotic arm inserts the sensor assembly into the ladle. These images are preprocessed sequentially, including grayscale conversion, contrast enhancement, binarization, and edge detection. Then, character segmentation and template matching or deep learning character recognition techniques are used to extract the ladle serial number string.

[0064] In one embodiment, to improve recognition robustness, the system performs voting fusion on the recognition results of multiple frames of images, and selects the recognition result with the highest frequency as the final number. Simultaneously, when the recognition confidence level falls below a certain threshold, the system prompts the on-site operator to manually input or supplement the number information using a handheld scanner.

[0065] Steel ladle identification information is obtained through an interface with the Manufacturing Execution System (MES). In the digital production environment of a steel plant, the MES typically maintains a complete ladle lifecycle file, including ladle number, ladle placement date, cumulative heat usage, and last repair record. This system communicates with the MES in real time via OPC UA, Modbus TCP, or direct database connection. Before or after each monitoring session, the system sends the currently identified ladle number to the MES, requesting a unique identifier, production status, usage history, and other auxiliary information for that ladle. This interface interaction method further verifies the accuracy of the optical character recognition results and obtains background data that cannot be directly obtained from the image.

[0066] The monitoring data is associated with the unique identifier of the corresponding ladle and stored accordingly. After identity verification, the system stores all data generated from this monitoring, including the 3D model, temperature distribution, anomaly markers, estimated residual thickness, slag adhesion area distribution, shooting time, and robotic arm insertion depth, in a structured format using the ladle's unique identifier as the primary key, in a local database or cloud server. Simultaneously, a monitoring record entry is generated, including key fields such as ladle number, monitoring time, operating team, and cumulative furnace usage update value, facilitating subsequent querying, statistics, and trend analysis.

[0067] For example, in the continuous casting workshop of a steel plant, ladle SP-0428 was monitored after its 187th heat. The system confirmed the number SP-0428 using visible light identification and obtained information from the MES interface, such as the ladle's cumulative usage of 186 heats and the last major overhaul three months prior. After monitoring, all data was stored under the identifier SP-0428_20260211_1432. Subsequent maintenance personnel can retrieve the ladle's historical thickness variation curves, erosion hotspot migration trends, and slag accumulation at any time, providing data support for developing personalized ladle repair strategies.

[0068] The display and alarm unit serves as the interface between the system and on-site operators, and mainly consists of an industrial display screen, an audible and visual alarm, and necessary voice broadcasting devices.

[0069] The system displays a real-time 3D point cloud model or unfolded cross-sectional view of the ladle lining with temperature pseudo-color overlay. Industrial displays typically use high-brightness LCD screens or touchscreen all-in-ones, installed on the ladle operating platform or in the central control room. The system renders the 3D point cloud model with temperature attributes into a pseudo-color image in real time, where temperatures are mapped sequentially from low to high to blue, green, yellow, orange, and red color ranges, allowing users to intuitively see the temperature distribution characteristics of different parts of the lining. Simultaneously, the system also supports unfolding the 3D model longitudinally or laterally into a 2D cross-sectional view, facilitating observation of the thickness and temperature correspondence of key components such as the ladle body, slag line, and bottom.

[0070] In one implementation, the display interface supports switching between multiple views, including an overall 3D perspective view, a magnified partial view, a thickness contour map, and a temperature heat map overlay view. Users can adjust the viewing angle arbitrarily by touching or dragging with a mouse, and choose whether to overlay auxiliary information such as slag-laden area markings and abnormal risk indicators.

[0071] Abnormal areas are highlighted. Once the status assessment unit outputs a severe erosion risk or abnormal slag adhesion, the display unit immediately overlays a prominent highlighted box, flashing arrow, or darkened area at the corresponding location. For example, areas with residual thickness below the safety threshold are marked with a flashing red outline, abnormal temperature hotspots are marked with flashing yellow dots, and areas with slag adhesion are covered with a semi-transparent orange mask. Through these visual cues, operators can quickly locate the problem area within seconds, improving emergency response efficiency.

[0072] The display shows the ladle number, cumulative usage count, current thinnest point thickness and location, and highest temperature point and location. A fixed information bar at the top of the interface refreshes the ladle's basic status in real time, including the ladle number, current heat batch, cumulative usage count, and monitoring time. The key indicators bar below highlights the current thinnest point thickness and its spatial coordinates, the highest temperature point value and location, and the maximum slag adhesion area, among other core parameters. All values ​​are rounded to one decimal place and are distinguished by different colors for normal, warning, and alarm states, facilitating quick assessment of the ladle's current health level.

[0073] In one possible implementation, when the system determines that there is an extremely high risk, the audible and visual alarm immediately activates a red warning light and emits an intermittent buzzing sound. Simultaneously, the voice broadcast device announces in synthesized voice, "Severe corrosion risk detected in the bottom area of ​​ladle SP-0428; the thinnest point thickness is only 32% of the design value. Please confirm and arrange for handling immediately." This multimodal alarm method effectively attracts the attention of on-site personnel, preventing them from missing crucial alarm information due to distraction.

[0074] This invention also provides a ladle condition monitoring method based on multi-source sensor fusion, comprising the following steps: Step 1: Steel ladle identification, obtaining the identity information of the steel ladle to be inspected; capturing the number image of the steel ladle through a visible light camera and identifying the steel ladle number, or retrieving the steel ladle number and historical usage data from the factory database; Step 2: Laser scanning. Drive the robotic arm to extend the 3D laser sensor probe into the ladle and perform a 360° full-coverage scan of the refractory lining surface of the ladle to collect high-density point cloud data to obtain the geometry of the lining. Step 3: Temperature measurement. Use an infrared thermal imaging sensor to obtain the temperature distribution map of the steel ladle lining surface, and obtain accurate and reliable temperature data through appropriate filtering and calibration. Step 4: Image acquisition. Use a visible light camera to acquire visible light images of the ladle lining. If necessary, use image analysis to identify color changes or slag adhesion on the lining surface. Step 5: Data fusion. The data processing module aligns and fuses the data from the above-mentioned sensor sources in time and space, associates the temperature data with the three-dimensional geometric model, and generates a comprehensive model that includes both lining thickness information and temperature information. This model is then combined with visible light images to corroborate suspicious areas. Step 6: Status determination. Analyze the fused model data and evaluate the status of the ladle lining, including calculating the residual thickness of refractory material in each part, detecting hot spots with excessive temperature, and identifying slag accumulation at the ladle opening or inner wall; determine whether there are any abnormalities based on the preset safety threshold, such as the thickness of the lining at a certain point being lower than the threshold or the temperature rising abnormally. Step 7: Output the results. The analyzed ladle status information is output through the display and alarm unit. The display interface presents a three-dimensional / two-dimensional status diagram of the ladle lining, marking weak areas and abnormal temperature areas. When there is a safety hazard, an audible and visual alarm is triggered to prompt the operator to take appropriate measures.

[0075] In step 5, a one-to-one correspondence is established between the temperature matrix collected by the infrared thermal imaging sensor and the three-dimensional laser point cloud. This includes calibrating the coordinate systems of the two types of sensors to ensure that each temperature pixel is mapped to the correct position on the three-dimensional point cloud model of the steel ladle lining, thereby obtaining a point cloud model with temperature attributes.

[0076] In step 6, when the highest temperature area is detected at the top opening of the ladle or the bottom outlet of the ladle, it is determined that the high temperature may be caused by molten steel residue rather than thinning of the lining. Therefore, the abnormal temperature point is ignored in the lining corrosion assessment. When it is found that the residual thickness of a certain lining area is lower than the preset value and the temperature at the corresponding location is higher than the warning threshold, it is comprehensively judged that the refractory material in that area has been severely deteriorated, and a warning message that it needs to be repaired and replaced immediately is output.

[0077] This invention achieves all-weather, multi-dimensional monitoring of the ladle's condition through a combination of hardware and software. The key innovation lies in using a robotic arm equipped with multi-source probes and triple protection (internal water-cooling circuit + high-temperature insulation layer + external gas purging curtain) to collect three types of data online within a ladle at nearly 1000°C without cooling. This is combined with specific location filtering rules for high-temperature interference from residual molten steel and cross-verification using thickness and temperature as dual indicators to determine severely corroded areas.

[0078] Example 1: like Figure 1 As shown, the hardware components of the high-temperature environment multi-source sensor fusion ladle condition monitoring system of the present invention include: a robotic arm 1 installed next to the ladle repair station, multi-source sensor probes 2, a data processing and control cabinet 3 in the main control room, and an operating terminal 4. Using this system, the lining thickness and temperature status of the ladle can be quickly measured and evaluated without human intervention each time it is turned over to the hot repair station.

[0079] The robotic arm 1 has 6 degrees of freedom and can extend into the interior of the ladle or move along its edge when the ladle is in the hot repair position.

[0080] The multi-source sensor probe 2 is fixed to the end of the robotic arm 1. The probe integrates a 360° lidar 2a, an infrared thermal imager 2b, and an industrial camera 2c.

[0081] The structure of the sensor cooling and protection device is as follows: Figure 2As shown: a copper pipe water-cooling channel 21 surrounding the sensor continuously circulates cooling water to remove heat; the sensor housing is covered with a layer of high-efficiency heat-insulating material 22, such as ceramic fiber, to block external radiant high temperatures; and an annular nozzle 23 connected to a compressed air source is also provided at the front end of the probe, which can form a high-speed air curtain outward in the direction of the sensor's field of view to prevent dust particles from entering the sensor window area. This triple protection design ensures that the probe 2 can still work stably in the internal environment of a steel ladle at a temperature close to approximately 1000°C.

[0082] The data processing and control cabinet 3 houses an industrial computer, which is connected to the robotic arm and various sensors via cables. Once the ladle enters the monitoring position, the system begins to operate automatically according to the software-defined process.

[0083] The industrial camera 2c captures images of the identification plaque on the side wall of the steel ladle. The computer runs an image recognition algorithm to extract the steel ladle number (such as image character recognition OCR technology) and retrieves the file information of the steel ladle corresponding to that number from the system database, including the number of times it has been used before and the last measurement data of the inner lining thickness.

[0084] Industrial computer-controlled robotic arm 1 performs scanning actions, such as... Figure 3 As shown in the flowchart: while the robotic arm rotates around the steel ladle axis to scan one circle, the LiDAR 2a rotates at high speed and emits a laser beam to measure the distance and acquire point cloud data of the entire inner wall and bottom of the steel ladle lining; this process can be completed in about 1 minute, and the obtained point cloud is filtered and stitched to reconstruct the three-dimensional shape model of the current steel ladle lining.

[0085] Infrared thermal imager 2b performs panoramic thermal imaging of the ladle's interior. The captured thermal images are corrected by software to obtain the temperature field distribution on the inner lining surface, which is then aligned and mapped with the coordinate system of the laser point cloud, assigning a corresponding temperature value to each point on the 3D model. Simultaneously, if large pieces of residual slag or flaked refractory bricks are present on the inner wall of the ladle, the visible light images acquired by camera 2c can be used to assist in identifying these anomalies: for example, detecting dark slag areas relative to the background color of the refractory material through image analysis.

[0086] During the data fusion phase, the industrial computer executes intelligent analysis algorithms to comprehensively analyze the aforementioned multi-source data.

[0087] For lining erosion assessment, the system compares the current thickness model measured by laser with the original design model or the last measurement model in the ladle history archive to calculate the wear amount at various points in the lining. Then, referring to the temperature information at the corresponding location, it focuses on areas that are both severely worn and have significantly increased temperature—these are usually dangerous areas where the lining is extremely thin and high temperatures have already penetrated from behind, requiring special attention.

[0088] The system automatically applies interference filtering rules: if certain temperature hotspots appear at the opening of the ladle (such as near the slag line) and are accompanied by bright molten steel residue in the visible light image, it is determined that the high temperature is not caused by the lining being too thin, but may be a false signal caused by the molten steel not being completely poured out, and the program removes it from the anomaly list.

[0089] For slag adhesion detection, the system compares the laser point cloud with a standard ladle model to identify any extra protrusions at the ladle opening. It then combines the reflectivity characteristics of these protrusions at the laser wavelength with the color of the visible light image to determine the presence of a thick layer of slag. If a large amount of slag is confirmed, the system will prompt for ladle cleaning and simultaneously mark these areas when calculating the lining thickness to avoid incorrectly including the slag layer in the refractory layer thickness.

[0090] All analysis results are displayed centrally on the interface of operating terminal 4, such as Figure 4 As shown: The left side of the interface displays a 3D point cloud model of the ladle lining, with different colors representing different temperature ranges and overlaid text labels indicating the current thickness of each measuring point. The right side of the interface lists monitoring summaries, including key indicators such as ladle number, number of uses, average residual thickness, thinnest point thickness and location, and highest surface temperature and location. If any indicator exceeds a threshold (e.g., the thinnest point thickness is below the specified value or the highest temperature exceeds the safe value), the corresponding item will be highlighted in red, triggering a system buzzer and flashing warning lights to alert on-site personnel. Operators can then decide whether to send the ladle for cold repair, replace the lining, or perform additional preheating before the next use. Monitoring data is also simultaneously uploaded to the plant's intranet database, providing data support for engineering technicians to analyze the lifespan of each ladle and develop maintenance plans.

[0091] As can be seen from the above embodiments, the multi-source sensor fusion ladle monitoring system of the present invention exhibits excellent performance in practical applications: it can complete the detection immediately while the ladle is still red-hot, without affecting the production rhythm; the fusion analysis effectively reduces the false alarm rate, issuing warnings only for truly dangerous situations; the entire process requires no manual intervention, greatly improving the safety and timeliness of ladle inspection.

[0092] This invention utilizes a comprehensive system employing multiple sensing methods, including laser 3D scanning, infrared thermal imaging, and visible light vision, to monitor the refractory lining of steel ladles from all angles. The system comprises a multi-sensor probe mounted on a robotic arm, along with its high-temperature protective device, and a data processing and control module. This module acquires the lining's geometry, temperature field distribution, and image information under conditions of high-temperature red-hot ladle and smoke interference. It then uses a data fusion algorithm to determine state parameters such as lining thickness, abnormal temperature areas, and slag adhesion. The system can detect weak areas or overheating signs in the ladle lining in real time and provide early warning signals, preventing accidents such as ladle leaks caused by excessive lining erosion or temperature runaway. Furthermore, this invention improves the accuracy and robustness of monitoring results through multi-source information cross-validation, achieving intelligent and digital management of the ladle's condition. Multiple field trials have verified that the system successfully issued early warnings for several ladle lining anomalies, significantly reducing safety risks in steelmaking production. The above provides a detailed description of the multi-source sensor fusion ladle condition monitoring system and method for high-temperature environments provided in this application.

[0093] The above description of the embodiments is only for the purpose of helping to understand the method and core idea of ​​this application; at the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0094] Certain terms are used in the specification and claims to refer to specific components. Those skilled in the art will understand that hardware manufacturers may use different names to refer to the same component. This specification and claims do not distinguish components based on differences in name, but rather on differences in function. The terms "comprising" and "including" used throughout the specification and claims are open-ended and should be interpreted as "comprising / including but not limited to". "Approximately" means that within an acceptable margin of error, those skilled in the art can solve the technical problem and substantially achieve the technical effect within a certain margin of error. The following descriptions in the specification are preferred embodiments for carrying out this application; however, these descriptions are for the purpose of illustrating the general principles of this application and are not intended to limit the scope of this application. The scope of protection of this application shall be determined by the appended claims.

[0095] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a product or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the product or system that includes said element.

[0096] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0097] The foregoing description illustrates and describes several preferred embodiments of this application. However, as previously stated, it should be understood that this application is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the application concept described herein through the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of this application should be within the protection scope of the appended claims.

Claims

1. A ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions, characterized in that, The high-temperature environment multi-source sensor fusion ladle condition monitoring system includes: The system includes a robotic arm, multi-source sensor components, cooling and protection devices, a data processing and control module, and a display and alarm unit. The robotic arm device is used to position the multi-source sensor assembly to a detection location inside a steel ladle that is in a high-temperature, red-hot state; The multi-source sensor assembly is installed at the end of the robotic arm device and includes a first imaging sensor, a second imaging sensor and a third imaging sensor, which are used to collect point cloud geometric data, surface temperature distribution data and visible light image data of the steel ladle refractory lining, respectively. The cooling and protection device covers the multi-source sensor assembly and includes an internal water-cooling circuit, a high-temperature insulation layer and an external gas purging structure, which is used to maintain the operating temperature of each sensor and keep the optical window clean in a high-temperature and dusty environment. The data processing and control module is connected to the multi-source sensor component and is used to perform spatiotemporal alignment and fusion processing on the three-source data to generate a three-dimensional point cloud model of the lining with temperature attributes, and output the ladle lining status information according to the preset judgment rules. The display and alarm unit is used to present the status information of the ladle lining and issue an alarm signal when an abnormality is detected.

2. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 1, characterized in that: The first imaging sensor is a three-dimensional laser sensor, used to collect point cloud geometric data of the refractory lining of the steel ladle; The second imaging sensor is an infrared thermal imaging sensor, used to collect surface temperature distribution data of the refractory lining of the steel ladle; The third imaging sensor is a visible light camera, used to collect visible light image data of the refractory lining of the steel ladle; The data processing and control module includes a multi-source data fusion unit, which is used for: The point cloud data of the three-dimensional laser sensor is used as the geometric reference coordinate system; The temperature pixel matrix of the infrared thermal imaging sensor is mapped to the point cloud coordinate system through pre-calibrated intrinsic and extrinsic parameters to form a temperature attribute point cloud. The image pixels of the visible light camera are mapped to the point cloud coordinate system through pre-calibrated intrinsic and extrinsic parameters to form an image attribute point cloud.

3. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 1, characterized in that: The internal water-cooling circuit is a cooling water pipe that is coiled around the multi-source sensor assembly; The high-temperature insulation layer is wrapped between the housing of the multi-source sensor assembly and the water-cooling circuit; The external gas purging structure includes an annular nozzle disposed at the front end of the multi-source sensor assembly, used to spray compressed air into the sensor's field of view to form an air curtain.

4. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 1, characterized in that, The data processing and control module further includes a status determination unit, which is used for: The residual thickness of each part of the steel ladle lining is calculated based on the temperature attribute points. Identify abnormal surface temperature regions based on the temperature attribute point cloud; The slag adhesion area is identified based on the image attribute point cloud.

5. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 4, characterized in that, The state determination unit includes an interference filtering subunit, which is used for: When the highest temperature zone is located in the ladle top edge opening area or the tapping area, the temperature anomaly is determined to be due to molten steel residue interference; the temperature influence of this zone is ignored in the thickness assessment. Based on the brightness characteristics of the image attribute point cloud, when there are bright residual molten steel image features in the top edge opening area or the steel outlet area, it is further confirmed that the high temperature is due to residual molten steel interference.

6. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 4, characterized in that, The status determination unit is also used to: determine that there is a serious risk of corrosion in a certain area when the residual thickness of a certain area is lower than the first threshold and the surface temperature of the area is higher than the second threshold; generate a corresponding abnormality mark and send it to the display and alarm unit.

7. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 1, characterized in that, The data processing and control module also includes a ladle identity management unit, which is used to: perform optical character recognition based on the ladle number image captured by the visible light camera; obtain ladle identification information through the interface with the manufacturing execution system; and associate and store the monitoring data with the unique identifier of the corresponding ladle.

8. The ladle condition monitoring system based on multi-source sensor fusion under high-temperature conditions according to claim 1, characterized in that, The display and alarm unit includes an industrial display screen and an audible and visual alarm, wherein the industrial display screen is used for: Real-time display of a 3D point cloud model or unfolded cross-sectional view of the ladle lining with temperature pseudo-color overlay; Highlight and mark abnormal areas; Displays the ladle number, cumulative number of uses, current thinnest point thickness and location, and highest temperature point and location.

9. A method for monitoring the condition of a steel ladle under high temperature conditions using multi-source sensor fusion, wherein monitoring is performed using the multi-source sensor fusion steel ladle condition monitoring system according to any one of claims 2-8, characterized in that... The method for monitoring the condition of a steel ladle under high-temperature conditions using multi-source sensor fusion includes the following steps: Step 1: Steel ladle identification, obtaining the identity information of the steel ladle to be inspected; capturing the number image of the steel ladle through a visible light camera and identifying the steel ladle number, or retrieving the steel ladle number and historical usage data from the factory database; Step 2: Laser scanning. Drive the robotic arm to extend the 3D laser sensor probe into the ladle and perform a 360° full-coverage scan of the refractory lining surface of the ladle to collect high-density point cloud data to obtain the geometry of the lining. Step 3: Temperature measurement. Use an infrared thermal imaging sensor to obtain the temperature distribution map of the steel ladle lining surface, and obtain accurate and reliable temperature data through appropriate filtering and calibration. Step 4: Image acquisition. Use a visible light camera to acquire visible light images of the ladle lining. Under specific conditions, image analysis can be used to identify color changes or slag adhesion on the lining surface. Step 5: Data fusion. The data processing module aligns and fuses the data from the above-mentioned sensor sources in time and space, associates the temperature data with the three-dimensional geometric model, and generates a comprehensive model that includes both lining thickness information and temperature information. This model is then combined with visible light images to corroborate suspicious areas. Step 6: Status determination. Analyze the fused model data and evaluate the status of the ladle lining, including calculating the residual thickness of refractory material in each part, detecting hot spots with excessive temperature, and identifying slag accumulation at the ladle opening or inner wall; determine whether there are any abnormalities based on the preset safety threshold, including the thickness of the lining at a certain point being lower than the threshold or the temperature rising abnormally. Step 7: Output the results. The analyzed ladle status information is output through the display and alarm unit. The display interface presents a three-dimensional / two-dimensional status diagram of the ladle lining, marking weak areas and abnormal temperature areas. When there is a safety hazard, an audible and visual alarm is triggered to prompt the operator to take appropriate measures.

10. The method for monitoring the condition of a steel ladle under high-temperature conditions using multi-source sensor fusion according to claim 9, characterized in that, In step 5, a one-to-one correspondence is established between the temperature matrix collected by the infrared thermal imaging sensor and the three-dimensional laser point cloud. This includes calibrating the coordinate systems of the two types of sensors to ensure that each temperature pixel is mapped to the correct position on the three-dimensional point cloud model of the steel ladle lining, thereby obtaining a point cloud model with temperature attributes. In step 6, when the highest temperature area is detected at the top opening of the ladle or the bottom outlet of the ladle, it is determined that the high temperature is caused by molten steel residue rather than thinning of the lining, and the abnormal temperature point is ignored in the lining corrosion assessment; when it is found that the residual thickness of a certain lining area is lower than the preset value and the temperature at the corresponding location is higher than the warning threshold, it is comprehensively judged that the refractory material in the area has been severely deteriorated, and a warning message that it needs to be repaired and replaced immediately is output.

Images