Intelligent analysis method, device and equipment for global three-dimensional deformation of boiler surface and storage medium

Abstract

The application discloses a boiler surface global three-dimensional deformation intelligent analysis method and device, equipment and a storage medium, and relates to the technical field of industrial detection. The method comprises the following steps: under the condition that multiple mobile robots cooperatively scan a boiler surface, acquiring three-dimensional point cloud data, measured data and robot attribute data, and acquiring multispectral reflection data of the boiler surface; performing registration processing on the three-dimensional point cloud data according to the robot attribute data to obtain registration data; constructing a three-dimensional topographic point cloud model of the boiler surface according to the registration data; correcting errors in the three-dimensional topographic point cloud model according to the measured data and the multispectral reflection data to obtain a three-dimensional topographic point cloud correction model; determining a deformation distribution nephogram of each region of the boiler surface according to the three-dimensional topographic point cloud correction model and a preset point cloud baseline; and analyzing a geometric deformation evolution trend of the deformation distribution nephogram to obtain an analysis result of the boiler surface.

Description

Technical Field

[0001] This application relates to the field of industrial testing technology, and in particular to intelligent analysis methods, devices, equipment and storage media for three-dimensional deformation of the entire surface of boilers. Background Technology

[0002] As core pressure-bearing equipment in thermal power, nuclear power, and industrial sectors, boilers inevitably undergo macroscopic or microscopic structural deformations during long-term operation due to factors such as thermal expansion and contraction, material creep, and fatigue accumulation. These deformations include overall expansion, localized bulges, dents, twisting, and even cracks. If these deformations are not detected and effectively controlled in a timely manner, they will seriously threaten the safety and economy of boiler operation. Therefore, it is necessary to monitor the three-dimensional deformation of the boiler surface.

[0003] Currently, boiler deformation monitoring uses non-contact laser displacement sensors to measure unidirectional or triaxial thermal expansion displacement at key points on the boiler. However, while existing methods can provide high-precision local data, they are limited by the number of sensors and their installation locations, making it impossible to directly capture the true morphology of the boiler surface. They also lack the ability to detect irregular deformations such as bulges and depressions, and large-scale coverage is costly. Therefore, how to perform intelligent three-dimensional deformation analysis of the boiler surface has become an urgent problem to be solved. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, equipment, and storage medium for intelligent analysis of full-area three-dimensional deformation of boiler surface, aiming to solve the technical problem of how to perform intelligent analysis of full-area three-dimensional deformation of boiler surface.

[0005] To achieve the above objectives, this application proposes a method for intelligent analysis of three-dimensional deformation of the entire boiler surface, the method comprising: With multiple mobile robots collaboratively scanning the boiler surface, three-dimensional point cloud data, measured data, and robot attribute data are acquired, and multispectral reflectance data of the boiler surface is also acquired. The 3D point cloud data is registered based on the robot attribute data to obtain registration data. A three-dimensional point cloud model of the boiler surface is constructed based on the registration data; Based on the measured data and the multispectral reflectance data, the errors in the three-dimensional topography point cloud model are corrected to obtain a corrected three-dimensional topography point cloud model. Based on the three-dimensional topographic point cloud correction model and the preset point cloud baseline, the deformation distribution cloud map of each region on the boiler surface is determined; The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface.

[0006] In one embodiment, the step of registering the 3D point cloud data based on the robot attribute data to obtain registered data includes: Based on the robot posture data and robot position data in the robot attribute data, the three-dimensional point cloud data is locally corrected to obtain locally corrected point cloud data. The local corrected point cloud data is matched with the three-dimensional coordinates of geometric reference points preset on the boiler surface to obtain the matching result; Based on the matching results, the local corrected point cloud data is globally iteratively registered using an iterative nearest point strategy to obtain the registered data in a preset coordinate system.

[0007] In one embodiment, the step of correcting the errors in the three-dimensional topography point cloud model based on the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topography point cloud model includes: Obtain a temperature distribution map; Based on the temperature, humidity, and air pressure data in the measured data, and combined with the local air temperature gradient of each region in the temperature distribution map, the laser propagation speed error in the three-dimensional topographic point cloud model is corrected to obtain the first correction result. Based on the multispectral reflectance data, the material type of each region on the boiler surface is identified, and combined with the surface temperature value of each region in the temperature distribution map, the material thermal expansion error of each region in the three-dimensional topography point cloud model is corrected to obtain the second correction result. Based on the vibration data in the measured data, the spatial ambiguity error in the three-dimensional topography point cloud model is corrected to obtain a third correction result; A three-dimensional topographic point cloud correction model is determined based on at least one of the first correction result, the second correction result, and the third correction result.

[0008] In one embodiment, the step of determining the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topographic point cloud correction model and the preset point cloud baseline includes: Using the spatial coordinates of each reference spatial point on the preset point cloud baseline as a reference, the difference between the spatial coordinates of the measured points in the neighborhood of the three-dimensional topography point cloud correction model under the spatial coordinates is calculated to obtain the three-dimensional geometric deviation value under the spatial coordinates. The three-dimensional geometric deviation values ​​are mapped onto the three-dimensional topographic point cloud correction model, and corresponding color levels are assigned according to the three-dimensional geometric deviation values ​​to obtain the deformation distribution cloud map of each region on the boiler surface.

[0009] In one embodiment, the step of analyzing the geometric deformation evolution trend of the deformation distribution cloud map to obtain the analysis results of the boiler surface includes: Obtain historical deformation distribution cloud maps; Based on the historical deformation distribution cloud map and the deformation distribution cloud map, the data are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results. The analysis results of the geometric deformation evolution trend of the predicted results are obtained by analyzing the surface of the boiler.

[0010] In one embodiment, the step of inputting the historical deformation distribution cloud map and the deformation distribution cloud map into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction result includes: Based on the historical deformation distribution cloud map and the deformation distribution cloud map, time series processing is performed to determine the deformation time series data of each spatial region. The deformation time series data includes the deformation rate and deformation acceleration of each spatial region. The deformation rate and deformation acceleration are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results.

[0011] In one embodiment, the step of analyzing the geometric deformation evolution trend of the predicted results to obtain the analysis results of the boiler surface includes: The geometric deformation evolution trend of the predicted results is analyzed based on a preset deformation safety threshold to obtain deformation mutation information, which includes deformation anomaly acceleration information and mutation risk information. Based on the deformation anomaly acceleration information and the mutation risk information, the analysis results of the boiler surface are obtained, including the location of the risk area, the risk type, and the warning level.

[0012] Furthermore, to achieve the above objectives, this application also proposes a boiler surface three-dimensional deformation intelligent analysis device, which includes: The acquisition module is used to acquire three-dimensional point cloud data, measured data, and robot attribute data when multiple mobile robots perform collaborative scanning of the boiler surface, and to acquire multispectral reflectance data of the boiler surface. The registration module is used to perform registration processing on the three-dimensional point cloud data according to the robot attribute data to obtain registration data; A construction module is used to construct a three-dimensional topographic point cloud model of the boiler surface based on the registration data; The correction module is used to correct the errors in the three-dimensional topography point cloud model based on the measured data and the multispectral reflectance data, so as to obtain a corrected three-dimensional topography point cloud model. The execution module is used to determine the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topography point cloud correction model and the preset point cloud baseline. The analysis module is used to analyze the geometric deformation evolution trend of the deformation distribution cloud map and obtain the analysis results of the boiler surface.

[0013] Furthermore, to achieve the above objectives, this application also proposes a boiler surface three-dimensional deformation intelligent analysis device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the boiler surface three-dimensional deformation intelligent analysis method as described above.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the intelligent analysis method for full-domain three-dimensional deformation of boiler surface as described above.

[0015] One or more technical solutions proposed in this application have at least the following technical effects: This embodiment proposes a method for intelligent analysis of three-dimensional deformation of the entire boiler surface. This method involves multiple mobile robots collaboratively scanning the boiler surface to acquire three-dimensional point cloud data, measured data, and robot attribute data, as well as multispectral reflectance data of the boiler surface. The three-dimensional point cloud data is then registered using the robot attribute data to obtain registered data. A three-dimensional topographic point cloud model of the boiler surface is constructed based on the registered data. Errors in the three-dimensional topographic point cloud model are corrected using the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topographic point cloud model. Based on the corrected three-dimensional topographic point cloud model and a preset point cloud baseline, a deformation distribution cloud map of each region of the boiler surface is determined. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface. This application utilizes multiple mobile robots to collaboratively scan the entire surface of a boiler, simultaneously acquiring various data. After registration, a three-dimensional topographic point cloud model of the boiler surface is constructed. Furthermore, the model is dynamically corrected by combining measured data and multispectral reflectance data to eliminate interference from factors such as temperature, vibration, and medium refraction on laser measurements. By comparing the corrected model with a preset point cloud baseline, a deformation distribution cloud map is generated. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface. This effectively identifies and quantifies subtle deformations such as local bulges and depressions, enabling intelligent analysis of deformation anomalies, early identification of potential structural failure risks, and improvement of boiler operation safety. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating an embodiment of the intelligent analysis method for full-area three-dimensional deformation of boiler surface in this application. Figure 2 This is a flowchart illustrating the overall process of the intelligent analysis method for full-area three-dimensional deformation of boiler surface in this application. Figure 3 This is a flowchart illustrating Embodiment 2 of the intelligent analysis method for full-area three-dimensional deformation of boiler surface provided in this application. Figure 4 This is a schematic diagram of the module structure of the intelligent analysis device for full-area three-dimensional deformation of boiler surface according to an embodiment of this application; Figure 5 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the intelligent analysis method for full-area three-dimensional deformation of boiler surface in this embodiment of the application.

[0019] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0021] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0022] The main solution of this application embodiment is as follows: When multiple mobile robots collaboratively scan the boiler surface, three-dimensional point cloud data, measured data, and robot attribute data are acquired, along with multispectral reflectance data of the boiler surface. The three-dimensional point cloud data is then registered based on the robot attribute data to obtain registration data. A three-dimensional topographic point cloud model of the boiler surface is constructed based on the registration data. Errors in the three-dimensional topographic point cloud model are corrected based on the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topographic point cloud model. The deformation distribution cloud map of each region on the boiler surface is determined based on the corrected three-dimensional topographic point cloud model and a preset point cloud baseline. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface.

[0023] In this embodiment, for ease of description, the following description will focus on the intelligent analysis device for identifying three-dimensional deformation of the entire boiler surface.

[0024] Existing technologies for boiler deformation monitoring employ non-contact laser displacement sensors to measure unidirectional or triaxial thermal expansion displacement at key points on the boiler. However, while these methods can provide high-precision local data, they are limited by the number of sensors and their installation locations, making it impossible to directly capture the true morphology of the boiler surface. They also lack the ability to detect irregular deformations such as bulges and dents, and large-scale coverage is costly.

[0025] This application provides a solution that uses multiple mobile robots to collaboratively scan the entire surface of a boiler, simultaneously acquiring various data. After registration, a three-dimensional topographic point cloud model of the boiler surface is constructed. Furthermore, the three-dimensional topographic point cloud model is dynamically corrected by combining measured data and multispectral reflectance data to eliminate interference from factors such as temperature, vibration, and medium refraction on laser measurements. By comparing the corrected model with a preset point cloud baseline, a deformation distribution cloud map is generated. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface. This effectively identifies and quantifies subtle deformations such as local bulges and depressions, enabling intelligent analysis of deformation anomalies, early identification of potential structural failure risks, and improvement of boiler operation safety.

[0026] Based on this, embodiments of this application provide a method for intelligent analysis of three-dimensional deformation of the entire boiler surface, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the intelligent analysis method for full-area three-dimensional deformation of boiler surface according to this application.

[0027] In this embodiment, the intelligent analysis method for full-area three-dimensional deformation of the boiler surface includes steps S10 to S40: Step S10: With multiple mobile robots collaboratively scanning the boiler surface, acquire three-dimensional point cloud data, measured data, and robot attribute data, and acquire multispectral reflectance data of the boiler surface. It should be noted that since a single fixed sensor has a limited coverage area and it is difficult to obtain the full three-dimensional topography of the boiler surface, multiple mobile robots can be used to scan the boiler surface collaboratively. The optimal collaborative scanning path and task can be planned for each robot. For example, the scanning frequency and density can be increased for key areas such as the furnace water-cooled wall and superheater header. At this time, three-dimensional point cloud data, measured data and robot attribute data can be obtained simultaneously, as well as multispectral reflectance data of the boiler surface.

[0028] Understandably, as a core pressure-bearing device, the boiler surface is the entire scannable area of ​​the boiler's outer wall. During long-term operation, the boiler surface will undergo macroscopic or microscopic structural deformations, including overall expansion, local bulging, dents, twisting, and even cracks, due to factors such as thermal expansion and contraction, material creep, and fatigue accumulation. Therefore, it is necessary to use the collected multi-source data to achieve full-area monitoring.

[0029] In a specific embodiment, the three-dimensional point cloud data is collected by a multispectral lidar module carried by each collaborative mobile robot. The multispectral lidar module uses a high-resolution LiDAR sensor (such as Velodyne VLP-32C or a customized multispectral LiDAR) with a point cloud measurement accuracy better than 0.5mm. It continuously scans the boiler surface at a set frequency (e.g., 10-20Hz) to obtain a set of discrete point coordinates characterizing the geometry of the boiler surface.

[0030] The measured data refers to the data collected by sensors, such as the precise temperature distribution map of the boiler surface obtained by an infrared thermal imager (such as FLIR T1020), with a temperature measurement accuracy better than ±1℃; the vibration acceleration data of the boiler structure and the robot itself collected by a multi-axis high-frequency vibration sensor (such as Kistler8763B triaxial accelerometer), with a sampling frequency of not less than 1kHz; and the local ambient temperature, humidity and air pressure data collected in real time by an ambient temperature, humidity and air pressure sensor (such as Bosch BME280).

[0031] The robot attribute data is output in real time by a high-precision inertial measurement unit (IMU) and odometry (such as the NovAtel OEM718DGNSS / IMU integrated navigation system), providing the robot's attitude and three-dimensional position information for point cloud registration and motion compensation, ensuring the accuracy and stability of the entire measurement system.

[0032] The multispectral reflectance data is obtained by a multispectral lidar module that transmits and receives laser signals at at least two different wavelengths (e.g., visible light, near-infrared, etc.), and simultaneously records the reflection intensity and spectral characteristics of the boiler surface material to different wavelength lasers, thus obtaining multispectral reflectance data. This data is used to characterize the differences in material properties, such as different types of metals, oxide scale, coatings, etc., exhibiting unique spectral characteristics in different bands.

[0033] Step S20: Perform registration processing on the 3D point cloud data according to the robot attribute data to obtain registration data; It should be noted that the registration data is a set of point clouds that accurately aligns 3D point cloud data from different robots, different perspectives, and different times to the same global coordinate system, which can eliminate spatial pose errors caused by the independent movement of multiple robots, changes in posture, and differences in scanning perspective.

[0034] In a specific embodiment, the three-dimensional point cloud data is locally corrected based on the robot posture data and robot position data in the robot attribute data to obtain locally corrected point cloud data. That is, the edge computing unit performs frame-by-frame distortion correction and motion compensation on the three-dimensional point cloud data collected by each robot in its respective scanning period based on the robot posture data and robot position data in the robot attribute data to obtain locally corrected point cloud data, so as to eliminate point cloud misalignment caused by the robot's own movement.

[0035] The local corrected point cloud data is matched with the three-dimensional coordinates of a geometric reference point preset on the boiler surface to obtain a matching result. That is, the local corrected point cloud data is matched with the three-dimensional coordinates of a geometric reference point (such as an artificially marked point or an inherent feature point of the boiler) preset on the boiler surface, and the initial transformation parameters between the local point cloud and the reference point are calculated to obtain the matching result.

[0036] Based on the matching results, the local correction point cloud data is globally iteratively registered using an iterative nearest point (ICP) strategy to obtain registered data in a preset coordinate system. Specifically, based on the matching results, the ICP strategy is used to globally iteratively register the local correction point cloud data of all robots in the multimodal data fusion and AI analysis module. By repeatedly optimizing the rotation and translation matrices between point clouds, the data converges to a preset error threshold (e.g., root mean square error less than 0.1 mm), resulting in high-precision registered data in a preset global coordinate system. At this point, the spatial accuracy of the registered data in any area on the boiler surface is better than the local accuracy of a single scan, providing a high-fidelity data source with spatial continuity and uniform error.

[0037] In one feasible implementation, step S20 may include steps A11 to A13: Step A11: Based on the robot posture data and robot position data in the robot attribute data, perform local correction on the three-dimensional point cloud data to obtain locally corrected point cloud data; It should be noted that the locally corrected point cloud data is a set of point clouds with high internal consistency in the robot's own local coordinate system after compensating for the point cloud distortion caused by the motion of a single robot. Its spatial accuracy is close to the theoretical optimal value of the original data of the scanner after motion compensation.

[0038] Understandably, robot posture data can be the robot's roll angle, pitch angle, and yaw angle, while robot position data can be the robot's translation coordinates in space, thereby eliminating point cloud stretching or overlapping caused by robot movement.

[0039] Step A12: Match the local correction point cloud data with the three-dimensional coordinates of the geometric reference points preset on the boiler surface to obtain the matching result; It should be noted that the matching result is the matching result between the local corrected point cloud and the reference point, which is used to coarsely align the local point cloud to the global reference coordinate system.

[0040] Understandably, the geometric reference points pre-set on the boiler surface are fixed feature points calibrated in advance by high-precision measuring instruments. They can provide reference benchmarks for point clouds from different robots and different perspectives, thereby overcoming the problem of initial alignment between multi-source point clouds that lack overlapping areas.

[0041] Step A13: Based on the matching result, the local corrected point cloud data is globally iteratively registered using an iterative nearest point strategy to obtain the registered data in a preset coordinate system.

[0042] It should be noted that the iterative nearest point strategy is an optimization algorithm that solves the optimal rigid body transformation by continuously searching for the nearest point pair between the source point cloud and the target point cloud and minimizing their weighted sum of squared distances. It can be extended to perform joint global optimization on all robot local point clouds, using the benchmark point matching result as the initial value to avoid getting trapped in local minima.

[0043] Understandably, the preset coordinate system is the original design coordinate system of the boiler or a global measurement coordinate system established through reference points. Its origin and axis are defined in accordance with the geometry and process requirements of the boiler. All registered point clouds are unified under this coordinate system, which can support robot operation path planning and defect localization.

[0044] Step S30: Construct a three-dimensional topographic point cloud model of the boiler surface based on the registration data; It should be noted that the three-dimensional topographic point cloud model is a three-dimensional model that characterizes the geometric shape of the boiler surface, including all topographic features such as smooth areas, weld bulges, corrosion depressions, and bulging deformations. It can be a point cloud structure with continuous and quantifiable spatial topological relationships to fully record the precise spatial location and size information of the geometry, concavity and convexity features, weld contours, and local defects of the boiler's outer or inner surface.

[0045] In a specific embodiment, the discrete point cloud in the registration data is converted into a continuous triangular mesh model, and the holes caused by the missing point cloud in the reconstructed mesh are repaired to eliminate reconstruction noise, thereby obtaining a three-dimensional morphological model of the boiler surface. The three-dimensional morphological model of the boiler surface has continuous, closed, and hole-free curved surface features, which can realistically restore all micro and macro morphological features of the boiler surface, such as smooth areas, weld bulges, corrosion depressions, and bulging deformations.

[0046] Step S40: Based on the measured data and the multispectral reflectance data, the errors in the three-dimensional topography point cloud model are corrected to obtain a corrected three-dimensional topography point cloud model. It should be noted that the three-dimensional topography point cloud correction model is a three-dimensional point cloud model after eliminating errors, so as to more realistically restore the actual geometric topography of the boiler surface under operating conditions.

[0047] In a specific embodiment, a temperature distribution map is obtained by continuously scanning the boiler surface with an infrared thermal imager deployed on a robot to obtain a high-resolution temperature distribution map in real time. The temperature distribution map can characterize the surface temperature field of each region with pixel-level accuracy.

[0048] Based on the measured temperature, humidity, and air pressure data, and combined with the local air temperature gradients of each region in the temperature distribution map, the laser propagation speed error in the three-dimensional topographic point cloud model is corrected to obtain the first correction result. That is, after the edge computing unit completes the initial filtering and compression, the laser propagation speed compensation module in the multi-source dynamic environment compensation calculates the actual propagation speed of the laser on different paths point by point, thereby correcting the ranging error caused by the change in air density in the three-dimensional topographic point cloud model and obtaining the first correction result. This effectively eliminates the laser refraction and speed deviation caused by the drastic temperature rise gradient between the high-temperature and low-temperature regions inside the boiler.

[0049] Based on the multispectral reflectance data, the material type of each region on the boiler surface is identified. Combined with the surface temperature values ​​of each region in the temperature distribution map, the thermal expansion error of the material in each region of the three-dimensional topography point cloud model is corrected to obtain the second correction result. That is, based on the multispectral reflectance data, the material type of each region on the boiler surface is accurately identified through a pre-trained spectral classification model. Combined with the accurate surface temperature values ​​of the corresponding regions provided by the infrared thermal imager, and referring to the baseline of the thermal expansion coefficient of each material in the boiler design parameters, the nonlinear thermal expansion amount of each region under actual working conditions is adaptively calculated by the material thermal expansion coefficient compensation module. This is used to correct the geometric distortion caused by thermal expansion and contraction in the three-dimensional topography point cloud model region by region to obtain the second correction result. This effectively eliminates the local deformation distortion caused by material differences and non-uniform temperature distribution, ensuring that even in areas of material phase change or slag coverage, the model can still truly reflect the original geometric topography of the boiler surface.

[0050] Based on the vibration data in the measured data, the spatial ambiguity error in the three-dimensional topographic point cloud model is corrected to obtain the third correction result. That is, the vibration acceleration data collected by the multi-axis high-frequency vibration sensor and the robot posture and motion trajectory data provided by the IMU / odometer are integrated. After synchronization and preliminary filtering in the edge computing unit, the vibration and robot motion compensation module separates the robot motion jitter and the low-frequency vibration of the boiler structure in real time. The spatial ambiguity and dynamic drift caused by relative displacement in the point cloud data are accurately compensated to obtain the third correction result. This effectively suppresses the point cloud dispersion effect caused by robot climbing vibration, robotic arm swing and boiler blower impact, so that the positioning accuracy of each spatial point is restored to the positioning accuracy level of the robot body.

[0051] Based on at least one of the first correction result, the second correction result, and the third correction result, a three-dimensional topographic point cloud correction model is determined, thereby eliminating the dynamic interference of laser path error, material thermal expansion error, and vibration and motion blur caused by the temperature field, and accurately restoring the true geometric shape of the boiler under operating conditions.

[0052] In one feasible implementation, step S40 may include steps B11-B15: Step B11: Obtain the temperature distribution map; It should be noted that the temperature distribution map is a two-dimensional temperature matrix generated by an infrared thermal imager deployed on a robot continuously scanning the boiler surface. Each pixel corresponds to the precise surface temperature value of a tiny area on the boiler surface, and its spatial resolution can reach the millimeter level. It can clearly characterize the non-uniform temperature field distribution formed by combustion, heat conduction and local slagging in the boiler from the cold state to the operating condition.

[0053] Understandably, the temperature distribution map allows all error corrections to be adaptively adjusted point-by-point or region-by-region based on the actual operating temperature field, avoiding the coarse errors caused by using the global average temperature.

[0054] Step B12: Based on the temperature, humidity and air pressure data in the measured data, and combined with the local air temperature gradient of each region in the temperature distribution map, the laser propagation speed error in the three-dimensional topographic point cloud model is corrected to obtain the first correction result. It should be noted that the first correction result is a three-dimensional topographic point cloud sub-model after laser propagation speed error correction. That is, by correcting the ranging deviation caused by the change in the propagation speed of the laser in the complex air medium around the boiler point by point, the spatial coordinates of each point in the point cloud are closer to the real geometric position.

[0055] Understandably, the temperature, humidity, and air pressure data in the measured data are global atmospheric parameters collected by environmental temperature, humidity, and air pressure sensors, representing the atmospheric state of the robot's environment. The local air temperature gradients in each region are extracted from the temperature distribution map, representing the differences in air temperature between different locations on the boiler surface, such as the abrupt temperature change between the near-fire and back-fire surfaces. The laser propagation speed error is due to the change in air density with temperature, humidity, and air pressure, causing the actual laser propagation speed to deviate from the speed of light in a vacuum, resulting in a systematic deviation in the lidar ranging value.

[0056] Step B13: Identify the material type of each region on the boiler surface based on the multispectral reflectance data, and combine the surface temperature values ​​of each region in the temperature distribution map to correct the material thermal expansion error of each region in the three-dimensional topography point cloud model, and obtain the second correction result; It should be noted that the second correction result is a three-dimensional topographic point cloud sub-model after adaptive correction of the material's thermal expansion coefficient. That is, by identifying different materials and correcting the amount of thermal expansion according to their actual temperature, the geometric distortion caused by thermal expansion and contraction in each region of the point cloud can be accurately restored.

[0057] It is understandable that the material type of each area on the boiler surface is obtained by identifying reflectance spectral data at different wavelengths. It can be carbon steel, alloy steel, oxide scale, slag, etc. The surface temperature value of each area in the temperature distribution map is the local true temperature directly measured by the infrared thermal imager. The material thermal expansion error in each area is due to the fact that different materials have different linear thermal expansion coefficients and the actual temperature is inconsistent with the reference temperature, which leads to a systematic deviation of non-uniform expansion of the boiler under operation.

[0058] Step B14: Based on the vibration data in the measured data, correct the spatial ambiguity error in the three-dimensional topographic point cloud model to obtain the third correction result; It should be noted that the third correction result is a three-dimensional topographic point cloud sub-model after vibration and motion compensation. That is, by suppressing the ambiguity and dynamic drift of the point cloud spatial position caused by the robot's own shaking and boiler structure vibration, the positioning accuracy of each spatial point in the point cloud is restored to the positioning accuracy level of the robot body.

[0059] Understandably, the vibration data in the measured data can characterize the instantaneous displacement and rotational disturbances caused by climbing walls, swinging of the robotic arm, boiler blowing, or structural resonance during the robot's movement. Spatial fuzziness error is due to the relative vibration or movement between the robot and the boiler surface during the lidar scanning process, which causes the same physical point in the point cloud to appear with positional dispersion, ghosting, or offset under multiple scans or different perspectives, thereby reducing the spatial consistency and clarity of the point cloud.

[0060] Step B15: Determine the three-dimensional topographic point cloud correction model based on at least one of the first correction result, the second correction result, and the third correction result.

[0061] Understandably, at least one of the first, second, and third correction results can be selected based on the actual working conditions to eliminate dynamic interference caused by laser propagation speed error, material thermal expansion error, and vibration ambiguity error using a three-dimensional topographic point cloud model.

[0062] Step S50: Determine the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topography point cloud correction model and the preset point cloud baseline. It should be noted that the deformation distribution cloud map of each region on the boiler surface is a spatial distribution map that visually represents the degree of deformation at each location on the boiler surface, and can display deformation features such as radial expansion, local depressions, bulges or twists.

[0063] Understandably, the preset point cloud baseline is an ideal three-dimensional point cloud model of the boiler under design conditions, representing the original geometric shape of the boiler surface under conditions without thermal stress and operational wear. By subtracting the three-dimensional topography point cloud correction model from the preset point cloud baseline, the cumulative deformation of the boiler wall under high temperature, pressure, and corrosion can be quantified.

[0064] In a specific embodiment, the spatial coordinates of each reference spatial point on the preset point cloud baseline are used as a reference. The difference between the spatial coordinates of the reference spatial points in the three-dimensional topography point cloud correction model and the spatial coordinates of the measured points in the neighborhood of the reference spatial coordinates is calculated to obtain the three-dimensional geometric deviation value under the spatial coordinates. The three-dimensional coordinates of each reference spatial point on the preset point cloud baseline are used as a reference. Through the high-precision point cloud registration and fusion algorithm in the multi-modal data fusion and AI analysis module, the three-dimensional topography point cloud correction model after multi-source dynamic environment compensation (including laser propagation speed, material thermal expansion coefficient and vibration error correction) is accurately aligned with the baseline. For each reference spatial point in the baseline, all measured points of the correction model in the neighborhood of the point coordinates (i.e., within a certain radius) are retrieved. The geometric coordinate estimate corresponding to the position is obtained by weighted average or nearest neighbor interpolation, and the difference between the estimate and the baseline point coordinates is calculated to obtain the three-dimensional geometric deviation value under the spatial coordinates.

[0065] The three-dimensional geometric deviation values ​​are mapped onto the three-dimensional topographic point cloud correction model, and corresponding color levels are assigned according to the three-dimensional geometric deviation values ​​to obtain the deformation distribution cloud map of each region on the boiler surface. That is, discrete deformation values ​​are mapped onto the surface of the entire three-dimensional topographic point cloud correction model through spatial interpolation to form a continuous deformation field. Corresponding color levels are assigned according to the numerical range of deformation amplitude. For example, areas with small deformation are marked as green, areas with medium deformation are marked as yellow, and areas with severe deformation are marked as red. The resulting deformation distribution cloud map of each region on the boiler surface can intuitively show the magnitude and distribution pattern of deformation in each region of the boiler surface, making it easy to quickly identify abnormal areas.

[0066] In one feasible implementation, step S50 may include steps C11-C12: Step C11: Using the spatial coordinates of each reference spatial point on the preset point cloud baseline as a reference, calculate the difference between the spatial coordinates of the measured points in the neighborhood of the three-dimensional topography point cloud correction model and the spatial coordinates of the measured points in the neighborhood of the spatial coordinates to obtain the three-dimensional geometric deviation value under the spatial coordinates. It should be noted that the three-dimensional geometric deviation value under the spatial coordinates is a three-dimensional vector, which represents the deformation amplitude and spatial direction of the boiler surface at the corresponding reference point.

[0067] It is understood that the spatial coordinates of the measured points in the neighborhood under the spatial coordinates are the set of three-dimensional coordinates of all measured points within a certain radius of the preset reference point coordinates in the three-dimensional topography point cloud correction model. Since the point cloud data has local sparsity and noise, a single nearest point cannot be directly used as the match. Therefore, it is necessary to estimate the real surface position corresponding to the reference point through multiple measured points in the neighborhood, so as to reduce the impact of random errors on deformation calculation.

[0068] Step C12: Map the three-dimensional geometric deviation value onto the three-dimensional topography point cloud correction model, and assign corresponding color levels according to the three-dimensional geometric deviation value to obtain the deformation distribution cloud map of each region on the boiler surface.

[0069] Understandably, the color levels are assigned based on the numerical range of deformation amplitude. For example, areas with small deformation are marked in green, areas with medium deformation are marked in yellow, and areas with severe deformation are marked in red. This results in a deformation distribution cloud map of each area on the boiler surface, which can intuitively show the magnitude and distribution pattern of deformation in each area of ​​the boiler surface, making it easy to quickly identify abnormal areas.

[0070] Step S60: Analyze the geometric deformation evolution trend of the deformation distribution cloud map to obtain the analysis results of the boiler surface.

[0071] It should be noted that the analysis results of the boiler surface are the analysis results of the geometric deformation evolution trend of each region of the boiler surface. For example, the current deformation amplitude and type of each region (such as overall expansion, local bulging, depression, and distortion), the spatial location and quantification level of abnormal regions, the deformation rate and its changing trend (acceleration or deceleration), the predicted value of key region deformation in the future (such as 48 to 72 hours), and the safety boundary level, etc.

[0072] In a specific embodiment, the deep learning deformation analysis and anomaly detection unit in the multimodal data fusion and AI analysis module can align the deformation distribution cloud maps generated from multiple monitoring periods in the time dimension and extract the sequence of three-dimensional geometric deviation values ​​at each spatial location changing over time. Using a temporal neural network model, such as a Long Short-Term Memory (LSTM) network or a Transformer, it can extract the complete sequence of three-dimensional geometric deviation values ​​at each reference spatial point from historical boiler deformation monitoring data. Simultaneously, each deformation sequence is labeled with a corresponding deformation mode label, forming supervised learning samples. Taking LSTM as an example, the input layer receives the feature vector for each time step, including features such as three-dimensional deviation components, deviation modulus, local curvature, and surface normal vector change. After passing through multiple stacked LSTM hidden layers, time-dependent features are extracted, and the fully connected classification layer outputs various deformation types. After training, the temporal neural network model can automatically identify different types of deformation patterns, such as uniform expansion, local bulging, deepening of indentation, and aggravation of distortion, and quantify their development rate and acceleration. At the same time, by comparing historical baseline data with current monitoring values, the temporal neural network model can detect abnormal areas where the deviation value exceeds the preset threshold or the evolution rate changes abruptly, and output their three-dimensional coordinates, severity level, and future trend prediction. This enables intelligent analysis of boiler deformation status from static quantification to dynamic evolution, and can identify potential risks such as structural instability, material fatigue, or local overload in advance. By feeding the analysis results back to the robot collaborative control and navigation platform, the system can be triggered to adaptively adjust the scanning frequency of key areas, thereby optimizing inspection efficiency while ensuring monitoring accuracy. This achieves early warning and proactive maintenance of boiler surface deformation, significantly reducing the risk of unplanned shutdowns and accidents.

[0073] At this time, the analysis results of the boiler surface can be updated in real time to the boiler digital twin platform through the 5G high-speed communication module, and the corresponding time-series neural network model can be updated synchronously. Based on the updated model, the digital twin platform performs high-precision finite element analysis (FEA) simulation in combination with the real-time operating parameters of the boiler (such as load, main steam pressure, furnace temperature, etc.) to evaluate the stress distribution, material fatigue accumulation and structural integrity of the deformation area, and dynamically calculates the safe boundary of operation in combination with AI deformation prediction and boiler design specifications. The platform sends deformation early warning information (audio-visual alarm, SMS notification, etc.) to the visualization and intelligent decision support interface, and provides intelligent decision suggestions, including operation optimization suggestions such as adjusting burner tilt angle, combustion ratio, and sootblower operation strategy, maintenance suggestions such as arranging inspection of specific areas or early local repair, and issuing emergency shutdown warnings when deformation or stress reaches the critical level.

[0074] At this point, the analysis results on the boiler surface allow for a direct view of the boiler's real-time three-dimensional morphology, deformation distribution, abnormal areas, predicted trends, and intelligent analysis reports. Furthermore, it enables remote monitoring of the operating status of multiple robots, adjustment of monitoring parameters, and issuance of commands. Based on the detected degree of deformation anomalies, prediction results, and changes in boiler operating conditions, the monitoring strategy is automatically adjusted. For example, when deformation in a certain area accelerates abnormally, the robot collaborative control and navigation platform immediately adjusts the scanning paths of the relevant robots, increasing the scanning frequency and density of that area to obtain denser deformation data for detailed analysis, thereby achieving a dynamic and intelligent adaptive monitoring cycle.

[0075] This embodiment proposes a method for intelligent analysis of three-dimensional deformation of the entire boiler surface. This method involves multiple mobile robots collaboratively scanning the boiler surface to acquire three-dimensional point cloud data, measured data, and robot attribute data, as well as multispectral reflectance data of the boiler surface. The three-dimensional point cloud data is then registered using the robot attribute data to obtain registered data. A three-dimensional topographic point cloud model of the boiler surface is constructed based on the registered data. Errors in the three-dimensional topographic point cloud model are corrected using the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topographic point cloud model. Based on the corrected three-dimensional topographic point cloud model and a preset point cloud baseline, a deformation distribution cloud map of each region of the boiler surface is determined. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface. This application utilizes multiple mobile robots to collaboratively scan the entire surface of a boiler, simultaneously acquiring various data. After registration, a three-dimensional topographic point cloud model of the boiler surface is constructed. Furthermore, the model is dynamically corrected by combining measured data and multispectral reflectance data to eliminate interference from factors such as temperature, vibration, and medium refraction on laser measurements. By comparing the corrected model with a preset point cloud baseline, a deformation distribution cloud map is generated. The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface. This effectively identifies and quantifies subtle deformations such as local bulges and depressions, enabling intelligent analysis of deformation anomalies, early identification of potential structural failure risks, and improvement of boiler operation safety.

[0076] It is understandable that, such as Figure 2 As shown, Figure 2This is the overall flowchart of the intelligent analysis method for full-area three-dimensional deformation of boiler surface in this application. The system performs initialization and baseline establishment, performs high-precision calibration of the overall boiler structure, establishes an initial three-dimensional deformation reference benchmark, and then proceeds to the collaborative scanning and multi-modal data acquisition stage. Multiple sensors such as lidar, ultrasonic phased array and thermal imaging work together to achieve full-area coverage of the boiler surface from point to surface and synchronous acquisition of multi-physics field data. Entering the edge computing and data processing stage, noise reduction, filtering and feature extraction are completed at the edge nodes on site to effectively reduce the data transmission pressure. After completing the underlying data preprocessing, through multi-modal data fusion and dynamic compensation, the multi-source data such as deformation, temperature and stress are aligned and unified in time and space, and the thermal deformation and mechanical vibration errors caused by boiler start-up and shutdown or operating condition fluctuations are compensated in real time. Based on this, the system performs three-dimensional deformation intelligent analysis and anomaly detection. It uses a pre-trained temporal neural network to identify deformation acceleration zones and abrupt change risk points, and combines historical trends to accurately locate potential failure areas. It uses digital twin model updates to achieve intelligent decision-making, synchronizes real-time monitoring data to the virtual boiler model, and deduces the deformation evolution path through model calculations. It automatically generates a graded response strategy from inspection reminders to shutdown maintenance. All operating statuses, deformation cloud maps, and early warning information are displayed in three dimensions through a remote monitoring platform and human-machine interface, making it easy for maintenance personnel to intuitively grasp the boiler's health status. At the same time, based on real-time boiler operating conditions (such as load and temperature changes) and deformation prediction results, it can feed back the end-of-line assessment information to the front-end data acquisition stage in a closed loop, dynamically adjusting the scanning frequency, sensor combination, and area focus, thereby constructing a self-learning and self-optimizing full-process cyclic monitoring system, significantly improving detection efficiency and early warning accuracy.

[0077] Based on the first embodiment of this application, in the second embodiment of this application, the same or similar content as the first embodiment can be referred to the above description, and will not be repeated hereafter.

[0078] In this embodiment, refer to Figure 3 , Figure 3 This is a flowchart illustrating Embodiment 2 of the intelligent analysis method for full-area three-dimensional deformation of boiler surface provided in this application. Step S60 specifically includes steps S61 to S63: Step S61: Obtain historical deformation distribution cloud map; In a specific embodiment, the historical deformation distribution cloud map is the deformation spectrum of the boiler in multiple past monitoring cycles, which can be used to extract the deformation evolution sequence of each spatial point. For example, when the historical deformation distribution cloud map of a certain area on the boiler surface shows that the deviation amplitude increases uniformly from +2mm to +5mm in three consecutive periods and the local curvature does not change abruptly, the sequence is marked as a uniform expansion sequence.

[0079] Step S62: Based on the historical deformation distribution cloud map and the deformation distribution cloud map, input them into the pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction result; It should be noted that the prediction results are the deformation prediction results of the boiler surface area for multiple consecutive monitoring periods in the future. The results are displayed in the form of a deformation evolution trend map (line graph or thermogram). Each area is marked with a different color to indicate the predicted deformation pattern, and a confidence level label is attached.

[0080] In a specific embodiment, time-series processing is performed based on the historical deformation distribution cloud map and the deformation distribution cloud map to determine the deformation time-series data of each spatial region. The deformation time-series data includes the deformation rate and deformation acceleration of each spatial region. The deformation rate and deformation acceleration are input into a pre-trained time-series neural network model to predict the deformation evolution trend of each region and obtain the prediction result. That is, based on the historical deformation distribution cloud map and the deformation distribution cloud map, the deformation amplitude sequence at different time points is extracted for each spatial region (e.g., a 5cm×5cm grid or each pixel point), and the deformation rate and deformation acceleration between adjacent monitoring periods are calculated to facilitate the early detection of early signs of deformation acceleration. At this time, the deformation rate and deformation acceleration of each spatial region are used as input features and input into the pre-trained time-series neural network model. By extracting the pattern of deformation rate and acceleration evolution over time, the time-series neural network model can output the prediction results of deformation amplitude, direction and deformation pattern category for each region in future consecutive monitoring periods, so that an early warning can be issued before the deformation reaches the danger threshold.

[0081] In one feasible implementation, step S62 may include steps D11-D12: Step D11: Perform time-series processing based on the historical deformation distribution cloud map and the deformation distribution cloud map to determine the deformation time-series data of each spatial region. The deformation time-series data includes the deformation rate and deformation acceleration of each spatial region. It should be noted that the deformation time series data is a sequence of deformation amplitude values ​​recorded sequentially at different monitoring time points for each spatial region, which can characterize the dynamic evolution record of the corresponding region in the time dimension.

[0082] It is understandable that the deformation rate of each spatial region represents the change in deformation amplitude per unit time, which can intuitively characterize the speed of deformation development. Deformation acceleration represents the rate of change of deformation rate, that is, whether the deformation is accelerating, decelerating or developing at a constant speed.

[0083] Additionally, it should be noted that time-series processing involves extracting and organizing the deformation amplitude data of each monitoring point in chronological order from the historical deformation distribution cloud map and the current deformation distribution cloud map of each spatial region, in order to quantify the dynamic evolution characteristics of deformation.

[0084] Step D12: Input the deformation rate and the deformation acceleration into the pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results.

[0085] It is understandable that the deformation evolution trend of each region is a regular change pattern of the geometric deformation of each region on the boiler surface over time, such as stable and slow growth, phased acceleration, sudden jumps, or periodic fluctuations.

[0086] Step S63: Analyze the geometric deformation evolution trend of the predicted results to obtain the analysis results of the boiler surface.

[0087] It is understandable that, based on the analysis results of the boiler surface, high-risk areas can be located before the deformation exceeds the safety threshold, which facilitates early warning. For example, an early warning can be issued when it is identified that the deformation range is expanding, the deformation direction is deflecting, or the deformation patterns of multiple adjacent areas are becoming consistent.

[0088] In a specific embodiment, the geometric deformation evolution trend of the predicted results is analyzed according to a preset deformation safety threshold to obtain deformation mutation information. The deformation mutation information includes deformation anomaly acceleration information and mutation risk information. Based on the deformation anomaly acceleration information and the mutation risk information, the analysis results of the boiler surface are obtained. The analysis results include the location of the risk area, the risk type, and the warning level. That is, by comparing the geometric evolution trend of the predicted results with the preset deformation safety threshold, deformation mutation information is obtained. Based on the deformation mutation information, the analysis results are output. The analysis results include the location of the risk area, the risk type, and the warning level, thereby accurately locating high-risk areas, effectively preventing sudden accidents caused by deformation mutations, and significantly enhancing the pre-control capability of boiler structural safety.

[0089] In one feasible implementation, step S63 may include steps E11-E12: Step E11: Analyze the geometric deformation evolution trend of the prediction result according to the preset deformation safety threshold to obtain deformation mutation information, which includes deformation anomaly acceleration information and mutation risk information. It should be noted that the deformation mutation information refers to the abnormal change characteristics identified by a preset safety threshold during the geometric deformation evolution of the boiler surface.

[0090] Understandably, abnormal acceleration information of deformation is a sudden increase in the growth rate of deformation per unit time or a continuous deviation from the normal evolution curve, while mutation risk information is a dangerous state with a high probability of structural mutation, such as a local bulge about to become unstable or a crack about to expand.

[0091] Step E12: Based on the deformation anomaly acceleration information and the mutation risk information, obtain the analysis results of the boiler surface. The analysis results include the location of the risk area, the risk type, and the warning level.

[0092] Understandably, the risk area location is a clearly marked spatial coordinate or range indicating the risk of abnormal acceleration or sudden change in deformation, used to accurately locate the location that requires key attention or intervention. The risk type is a label obtained by classifying the deformation changes exhibited by the risk area on the boiler surface, used to identify specific failure modes, such as plastic instability, fatigue accumulation, high-temperature creep, etc. The warning level is a label obtained by classifying the degree of danger posed by a specific risk area on the boiler surface, used to match differentiated response actions.

[0093] This embodiment proposes a comprehensive three-dimensional intelligent analysis method for boiler surface deformation. It acquires historical deformation distribution cloud maps; based on these historical and current deformation distribution cloud maps, it inputs them into a pre-trained temporal neural network model to predict the deformation evolution trend of each region, obtaining prediction results; and analyzes the geometric deformation evolution trend of the prediction results to obtain the analysis results for the boiler surface. This application, by acquiring historical and current deformation distribution cloud maps of the boiler surface and inputting them into a pre-trained temporal neural network model to predict the deformation evolution trend of each region, and then analyzing the geometric deformation evolution trend in the prediction results to obtain the analysis results for the boiler surface, can identify potential abnormal acceleration areas and sudden change risks in advance, thereby significantly improving the accuracy and timeliness of early warning, minimizing sudden structural failures, and reducing maintenance costs and safety risks.

[0094] This application also provides a smart analysis device for three-dimensional deformation of the entire boiler surface. Please refer to [reference needed]. Figure 4 The intelligent analysis device for full-area three-dimensional deformation of the boiler surface includes: The acquisition module 10 is used to acquire three-dimensional point cloud data, measured data, and robot attribute data when multiple mobile robots perform collaborative scanning on the boiler surface, and to acquire multispectral reflectance data of the boiler surface. Registration module 20 is used to perform registration processing on the three-dimensional point cloud data according to the robot attribute data to obtain registration data; The construction module 30 is used to construct a three-dimensional topographic point cloud model of the boiler surface based on the registration data. The correction module 40 is used to correct the errors in the three-dimensional topography point cloud model based on the measured data and the multispectral reflectance data, so as to obtain a corrected three-dimensional topography point cloud model. The execution module 50 is used to determine the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topography point cloud correction model and the preset point cloud baseline. The analysis module 60 is used to analyze the geometric deformation evolution trend of the deformation distribution cloud map and obtain the analysis results of the boiler surface.

[0095] The registration module 20 is further configured to perform local correction on the three-dimensional point cloud data based on the robot posture data and robot position data in the robot attribute data, to obtain locally corrected point cloud data; The local corrected point cloud data is matched with the three-dimensional coordinates of geometric reference points preset on the boiler surface to obtain the matching result; Based on the matching results, the local corrected point cloud data is globally iteratively registered using an iterative nearest point strategy to obtain the registered data in a preset coordinate system.

[0096] The correction module 40 is also used to acquire a temperature distribution map; Based on the temperature, humidity, and air pressure data in the measured data, and combined with the local air temperature gradient of each region in the temperature distribution map, the laser propagation speed error in the three-dimensional topographic point cloud model is corrected to obtain the first correction result. Based on the multispectral reflectance data, the material type of each region on the boiler surface is identified, and combined with the surface temperature value of each region in the temperature distribution map, the material thermal expansion error of each region in the three-dimensional topography point cloud model is corrected to obtain the second correction result. Based on the vibration data in the measured data, the spatial ambiguity error in the three-dimensional topography point cloud model is corrected to obtain a third correction result; A three-dimensional topographic point cloud correction model is determined based on at least one of the first correction result, the second correction result, and the third correction result.

[0097] The execution module 50 is further configured to use the spatial coordinates of each reference spatial point on the preset point cloud baseline as a reference, and perform difference calculation with the spatial coordinates of the measured points in the neighborhood of the three-dimensional topography point cloud correction model under the spatial coordinates to obtain the three-dimensional geometric deviation value under the spatial coordinates. The three-dimensional geometric deviation values ​​are mapped onto the three-dimensional topography point cloud correction model, and corresponding color levels are assigned according to the three-dimensional geometric deviation values ​​to obtain the deformation distribution cloud map of each region on the boiler surface.

[0098] The analysis module 60 is also used to obtain historical deformation distribution cloud maps; Based on the historical deformation distribution cloud map and the deformation distribution cloud map, the data are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results. The analysis results of the geometric deformation evolution trend of the predicted results are obtained by analyzing the surface of the boiler.

[0099] The analysis module 60 is further configured to perform time-series processing based on the historical deformation distribution cloud map and the deformation distribution cloud map to determine the deformation time-series data of each spatial region, wherein the deformation time-series data includes the deformation rate and deformation acceleration of each spatial region; The deformation rate and deformation acceleration are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results.

[0100] The analysis module 60 is also used to analyze the geometric deformation evolution trend of the prediction result according to the preset deformation safety threshold, and obtain deformation mutation information, which includes deformation anomaly acceleration information and mutation risk information. Based on the deformation anomaly acceleration information and the mutation risk information, the analysis results of the boiler surface are obtained, including the location of the risk area, the risk type, and the warning level.

[0101] The intelligent analysis device for full-area three-dimensional deformation of boiler surface provided in this application adopts the intelligent analysis method for full-area three-dimensional deformation of boiler surface in the above embodiments, which can solve the technical problem of how to perform intelligent analysis of full-area three-dimensional deformation of boiler surface. Compared with the prior art, the beneficial effects of the intelligent analysis device for full-area three-dimensional deformation of boiler surface provided in this application are the same as the beneficial effects of the intelligent analysis method for full-area three-dimensional deformation of boiler surface provided in the above embodiments, and other technical features in the intelligent analysis device for full-area three-dimensional deformation of boiler surface are the same as the features disclosed in the method of the above embodiments, and will not be repeated here.

[0102] This application provides a boiler surface three-dimensional deformation intelligent analysis device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the boiler surface three-dimensional deformation intelligent analysis method in the above embodiment 1.

[0103] The following is for reference. Figure 5This document illustrates a structural schematic diagram of a boiler surface three-dimensional deformation intelligent analysis device suitable for implementing embodiments of this application. The boiler surface three-dimensional deformation intelligent analysis device in this application embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PMPs (Portable Media Players), and vehicle terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The boiler surface three-dimensional deformation intelligent analysis device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0104] like Figure 5 As shown, the intelligent analysis device for the full-area three-dimensional deformation of the boiler surface may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in ROM (Read Only Memory) 1002 or the program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the intelligent analysis device for the full-area three-dimensional deformation of the boiler surface. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the boiler surface full-area three-dimensional deformation intelligent analysis device to wirelessly or wiredly communicate with other devices to exchange data. Although the figure shows a boiler surface full-area three-dimensional deformation intelligent analysis device with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0105] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0106] The intelligent analysis device for full-area three-dimensional deformation of boiler surface provided in this application adopts the intelligent analysis method for full-area three-dimensional deformation of boiler surface in the above embodiments, which can solve the technical problem of how to perform intelligent analysis of full-area three-dimensional deformation of boiler surface. Compared with the prior art, the beneficial effects of the intelligent analysis device for full-area three-dimensional deformation of boiler surface provided in this application are the same as the beneficial effects of the intelligent analysis method for full-area three-dimensional deformation of boiler surface provided in the above embodiments, and other technical features in the intelligent analysis device for full-area three-dimensional deformation of boiler surface are the same as the features disclosed in the method of the previous embodiment, and will not be repeated here.

[0107] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0108] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0109] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the intelligent analysis method for full-domain three-dimensional deformation of boiler surface in the above embodiments.

[0110] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0111] The aforementioned computer-readable storage medium may be included in the intelligent analysis device for full-area three-dimensional deformation of the boiler surface; or it may exist independently and not be assembled into the intelligent analysis device for full-area three-dimensional deformation of the boiler surface.

[0112] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the intelligent analysis device for full-domain three-dimensional deformation of the boiler surface, the intelligent analysis device for full-domain three-dimensional deformation of the boiler surface performs the following: acquiring three-dimensional point cloud data, measured data, and robot attribute data, and acquiring multispectral reflectance data of the boiler surface, under the condition that multiple mobile robots are collaboratively scanning the boiler surface; performing registration processing on the three-dimensional point cloud data according to the robot attribute data to obtain registration data; constructing a three-dimensional topographic point cloud model of the boiler surface based on the registration data; correcting errors in the three-dimensional topographic point cloud model based on the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topographic point cloud model; determining the deformation distribution cloud map of each region of the boiler surface based on the corrected three-dimensional topographic point cloud model and a preset point cloud baseline; and analyzing the geometric deformation evolution trend of the deformation distribution cloud map to obtain the analysis results of the boiler surface.

[0113] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0114] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0115] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0116] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described intelligent analysis method for full-area three-dimensional deformation of boiler surfaces. This solves the technical problem of how to perform intelligent analysis of full-area three-dimensional deformation of boiler surfaces. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the intelligent analysis method for full-area three-dimensional deformation of boiler surfaces provided in the above embodiments, and will not be repeated here.

[0117] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for intelligent analysis of three-dimensional deformation across the entire surface of a boiler, characterized in that, The method includes: With multiple mobile robots collaboratively scanning the boiler surface, three-dimensional point cloud data, measured data, and robot attribute data are acquired, and multispectral reflectance data of the boiler surface is also acquired. The 3D point cloud data is registered based on the robot attribute data to obtain registration data. A three-dimensional point cloud model of the boiler surface is constructed based on the registration data; Based on the measured data and the multispectral reflectance data, the errors in the three-dimensional topography point cloud model are corrected to obtain a corrected three-dimensional topography point cloud model. Based on the three-dimensional topographic point cloud correction model and the preset point cloud baseline, the deformation distribution cloud map of each region on the boiler surface is determined; The geometric deformation evolution trend of the deformation distribution cloud map is analyzed to obtain the analysis results of the boiler surface.

2. The method as described in claim 1, characterized in that, The step of registering the 3D point cloud data based on the robot attribute data to obtain the registered data includes: Based on the robot posture data and robot position data in the robot attribute data, the three-dimensional point cloud data is locally corrected to obtain locally corrected point cloud data. The local corrected point cloud data is matched with the three-dimensional coordinates of geometric reference points preset on the boiler surface to obtain the matching result; Based on the matching results, the local corrected point cloud data is globally iteratively registered using an iterative nearest point strategy to obtain the registered data in a preset coordinate system.

3. The method as described in claim 1, characterized in that, The step of correcting the errors in the three-dimensional topography point cloud model based on the measured data and the multispectral reflectance data to obtain a corrected three-dimensional topography point cloud model includes: Obtain a temperature distribution map; Based on the temperature, humidity, and air pressure data in the measured data, and combined with the local air temperature gradient of each region in the temperature distribution map, the laser propagation speed error in the three-dimensional topographic point cloud model is corrected to obtain the first correction result. Based on the multispectral reflectance data, the material type of each region on the boiler surface is identified, and combined with the surface temperature value of each region in the temperature distribution map, the material thermal expansion error of each region in the three-dimensional topography point cloud model is corrected to obtain the second correction result. Based on the vibration data in the measured data, the spatial ambiguity error in the three-dimensional topography point cloud model is corrected to obtain a third correction result; A three-dimensional topographic point cloud correction model is determined based on at least one of the first correction result, the second correction result, and the third correction result.

4. The method as described in claim 1, characterized in that, The step of determining the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topography point cloud correction model and the preset point cloud baseline includes: Using the spatial coordinates of each reference spatial point on the preset point cloud baseline as a reference, the difference between the spatial coordinates of the measured points in the neighborhood of the three-dimensional topography point cloud correction model under the spatial coordinates is calculated to obtain the three-dimensional geometric deviation value under the spatial coordinates. The three-dimensional geometric deviation values ​​are mapped onto the three-dimensional topographic point cloud correction model, and corresponding color levels are assigned according to the three-dimensional geometric deviation values ​​to obtain the deformation distribution cloud map of each region on the boiler surface.

5. The method as described in claim 1, characterized in that, The step of analyzing the geometric deformation evolution trend of the deformation distribution cloud map to obtain the analysis results of the boiler surface includes: Obtain historical deformation distribution cloud maps; Based on the historical deformation distribution cloud map and the deformation distribution cloud map, the data are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results. The analysis results of the geometric deformation evolution trend of the predicted results are obtained by analyzing the surface of the boiler.

6. The method as described in claim 5, characterized in that, The step of inputting the historical deformation distribution cloud map and the deformation distribution cloud map into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction result includes: Based on the historical deformation distribution cloud map and the deformation distribution cloud map, time series processing is performed to determine the deformation time series data of each spatial region. The deformation time series data includes the deformation rate and deformation acceleration of each spatial region. The deformation rate and deformation acceleration are input into a pre-trained temporal neural network model to predict the deformation evolution trend of each region and obtain the prediction results.

7. The method as described in claim 5, characterized in that, The step of analyzing the geometric deformation evolution trend of the predicted results to obtain the analysis results of the boiler surface includes: The geometric deformation evolution trend of the predicted results is analyzed based on a preset deformation safety threshold to obtain deformation mutation information, which includes deformation anomaly acceleration information and mutation risk information. Based on the deformation anomaly acceleration information and the mutation risk information, the analysis results of the boiler surface are obtained, including the location of the risk area, the risk type, and the warning level.

8. A smart analysis device for three-dimensional deformation of the entire surface of a boiler, characterized in that, The device includes: The acquisition module is used to acquire three-dimensional point cloud data, measured data, and robot attribute data when multiple mobile robots perform collaborative scanning of the boiler surface, and to acquire multispectral reflectance data of the boiler surface. The registration module is used to perform registration processing on the three-dimensional point cloud data according to the robot attribute data to obtain registration data; A construction module is used to construct a three-dimensional topographic point cloud model of the boiler surface based on the registration data; The correction module is used to correct the errors in the three-dimensional topography point cloud model based on the measured data and the multispectral reflectance data, so as to obtain a corrected three-dimensional topography point cloud model. The execution module is used to determine the deformation distribution cloud map of each region on the boiler surface based on the three-dimensional topography point cloud correction model and the preset point cloud baseline. The analysis module is used to analyze the geometric deformation evolution trend of the deformation distribution cloud map and obtain the analysis results of the boiler surface.

9. A smart analysis device for three-dimensional deformation of the entire surface of a boiler, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the intelligent analysis method for full-domain three-dimensional deformation of boiler surface as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the intelligent analysis method for full-area three-dimensional deformation of boiler surface as described in any one of claims 1 to 7.

Images