BIM-based thermal power plant metal welding visual management method and system

Abstract

The application provides a BIM-based thermal power plant metal welding visualization management method and system, relates to the technical field of welding, and comprises the following steps: acquiring metal welding material information for metal welding; establishing a pixel granularity mapping rule, inputting metal material types, metal material porosity densities and welding surface roughness into the pixel granularity mapping rule, and determining a visualization pixel processing granularity; collecting an initial welding visualization image set of a metal welding process, connecting a BIM visualization rendering engine platform, performing pixel grid granularity processing on the initial welding visualization image set according to the visualization pixel processing granularity, and obtaining a first welding visualization image set; performing welding defect detection on the first welding visualization image set, and outputting a welding visualization defect image. The application can solve the technical problem of poor welding management quality in the prior art and achieve the technical effect of improving the welding management quality.

Description

Technical Field

[0001] This application relates to the field of welding technology, and in particular to a BIM-based visual management method and system for metal welding in thermal power plants. Background Technology

[0002] With the continuous expansion of the construction and operation scale of thermal power plants, welding technology occupies a core position in equipment manufacturing and installation. In particular, the welding quality of pressure vessels, pipelines, boilers and large metal components is directly related to the safety and stability of power plant operation.

[0003] Currently, existing welding quality inspection methods mainly rely on manual non-destructive testing techniques, such as ultrasonic testing, X-ray testing, or magnetic particle testing. While these methods offer a certain level of accuracy in defect location and assessment, they generally suffer from several shortcomings. First, these inspection methods are often performed after welding is complete, making it difficult to detect early signs of defects during the welding process, potentially leading to rework or other quality risks. Second, manual inspection depends on the experience and skill level of the operators; different inspectors may have varying degrees of defect identification, affecting the consistency and objectivity of the results. Furthermore, traditional methods lack deep integration with digital platforms such as BIM, failing to directly map inspection results onto 3D models for visualization. This results in a fragmented welding quality management process, hindering the formation of full lifecycle quality traceability and dynamic optimization control.

[0004] In summary, existing technologies suffer from several technical problems. These include a lack of diverse welding quality monitoring methods, a lack of dynamic integration analysis of material properties and the welding process, and a lack of visualization integration with BIM platforms. Consequently, welding defects are not detected in a timely manner, and the detection results are not uniform or intuitive. These issues further affect the welding quality control level of key equipment in thermal power plants and the overall safety and reliability of their operation. Summary of the Invention

[0005] The purpose of this application is to provide a BIM-based visual management method and system for metal welding in thermal power plants, in order to solve the technical problems in the existing technology, such as the lack of a single welding quality monitoring method, the lack of dynamic integration analysis of material properties and welding process, and the lack of visualization integration with the BIM platform, which leads to untimely detection of welding defects, inconsistent and unintuitive detection results, and further affects the welding quality control level of key equipment in thermal power plants and the safety and reliability of overall operation.

[0006] In view of the above problems, this application provides a BIM-based visual management method and system for metal welding in thermal power plants.

[0007] Firstly, this application provides a BIM-based visual management method for metal welding in thermal power plants, implemented through a BIM-based visual management system for metal welding in thermal power plants. The method includes: acquiring information on metal welding materials used for metal welding; extracting the metal material type, metal porosity, and welding surface roughness from the metal welding material information; establishing pixel granularity mapping rules; inputting the metal material type, metal porosity, and welding surface roughness into the pixel granularity mapping rules to determine the visual pixel processing granularity; collecting an initial set of visual welding images of the metal welding process; connecting to a BIM visualization rendering engine platform; performing pixel mesh granularity processing on the initial set of visual welding images according to the visual pixel processing granularity to obtain a first set of visual welding images; and performing welding defect detection on the first set of visual welding images to output a visual welding defect image.

[0008] Preferably, the BIM-based visual management method for metal welding in thermal power plants further includes: determining whether the materials used for metal welding are the same metal material; if the materials used for metal welding are the same metal material, determining whether a transition metal material is used in the metal welding process; if a transition metal material is used, extracting the transition metal material type, transition metal porosity, and transition metal surface roughness of the transition metal material; inputting the transition metal material type, transition metal porosity, and transition metal surface roughness into the pixel granularity mapping rule to determine the transition visualization pixel processing granularity; and weightedly fusing the transition visualization pixel processing granularity and the visualization pixel processing granularity to output the fused visualization pixel processing granularity.

[0009] Preferably, the BIM-based visual management method for metal welding in thermal power plants further includes: if the materials used for metal welding are different metal materials, extracting the first metal welding material information corresponding to the first end welding metal and the second metal welding material information corresponding to the second end welding metal; the pixel granularity mapping rule obtaining the first visual pixel processing granularity and the second visual pixel processing granularity based on the first metal welding material information and the second metal welding material information respectively; and weightedly fusing the first visual pixel processing granularity and the second visual pixel processing granularity to output the fused visual pixel processing granularity.

[0010] Preferably, the BIM-based visual management method for metal welding in thermal power plants further includes: if the materials used for metal welding are different metal materials and the metal welding process does not use transitional metal materials, the first visual pixel processing granularity and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity; if the materials used for metal welding are different metal materials and the metal welding process uses transitional metal materials, the transitional visual pixel processing granularity, the first visual pixel processing granularity, and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity.

[0011] Preferably, the BIM-based visual management method for metal welding in thermal power plants further includes: constructing a metal material sample library, including average porosity samples of metal materials under different metal material types and average surface roughness samples of metal materials; collecting welding defect image samples from the metal material sample library under different metal welding conditions, scoring the defect display quality of the welding defect image samples at different pixel granularities, and obtaining a pixel granularity solution set of the welding defect image samples that meets a preset defect display quality threshold; constructing a defect-granularity association knowledge base based on the welding defect image samples and the pixel granularity solution set; assigning weight coefficients to the different metal material types, average porosity samples of metal materials, and average surface roughness samples of metal materials to construct a first weight coefficient; and learning and updating the defect-granularity association knowledge base with the first weight coefficient to obtain pixel granularity mapping rules.

[0012] Preferably, the BIM-based visual management method for metal welding in thermal power plants further includes: initializing a random forest regression model, including the number of random trees, the depth of the random trees, and the number of node levels of the random trees; encoding the samples of different metal material types, average porosity of metal materials, and average roughness of metal surfaces as input to initialize the training set of the random forest regression model, and training the model using the pixel-level solution set as supervision labels to obtain a trained random forest regression model; extracting the learning weight coefficients of the trained random forest regression model; comparing the learning weight coefficients with the first weight coefficients, and updating the model based on the weight coefficient deviation when the accuracy of the trained random forest regression model is less than a preset threshold to obtain a converged random forest regression model; and outputting pixel-level mapping rules based on the converged random forest regression model.

[0013] Preferably, the BIM-based metal welding visualization management method for thermal power plants further includes: establishing a connection with the BIM visualization rendering engine platform through an application programming interface (API) and issuing visualization pixel processing commands; the BIM visualization rendering engine platform performing coordinate transformation on the acquired initial welding visualization image set in the BIM model coordinate system based on the visualization pixel processing commands to obtain a BIM welding visualization mapping image set; and performing pixel grid granular processing on the BIM welding visualization mapping image set according to the visualization pixel processing granularity to obtain a first welding visualization image set.

[0014] Preferably, the BIM-based metal welding visualization management method for thermal power plants further includes: the BIM visualization rendering engine platform dividing the BIM welding visualization mapping image set into a grid according to the parametric grid components at the visualization pixel processing granularity, and outputting a first welding visualization image set; wherein, each welding visualization image in the first welding visualization image set includes a set of grid pixel blocks corresponding to the pixel grid granularity, and the set of grid pixel blocks is stored according to the index position.

[0015] Preferably, the BIM-based metal welding visualization management method for thermal power plants further includes: extracting feature values ​​of each welding visualization image in the first welding visualization image set with respect to grid pixel blocks, wherein the feature values ​​include average grayscale value, average color value, gradient contrast, and standard deviation; calculating the defect probability of each grid pixel block based on the feature values, marking grid pixel blocks with defect probabilities greater than a preset defect probability, and performing color rendering and marking on the BIM visualization rendering engine platform to output a welding visualization defect image.

[0016] Secondly, this application also provides a BIM-based metal welding visualization management system for thermal power plants, used to execute the BIM-based metal welding visualization management method for thermal power plants as described in the first aspect, including: a metal welding material information acquisition module, used to acquire metal welding material information for metal welding, and extract the metal material type, metal porosity, and welding surface roughness of the metal welding material information; a visualization pixel processing granularity determination module, used to establish pixel granularity mapping rules, inputting the metal material type, metal porosity, and welding surface roughness into the pixel granularity mapping rules to determine the visualization pixel processing granularity; a first welding visualization image set acquisition module, used to collect an initial welding visualization image set of the metal welding process, connect to a BIM visualization rendering engine platform, and perform pixel grid granularity processing on the initial welding visualization image set according to the visualization pixel processing granularity to obtain a first welding visualization image set; and a welding visualization defect image output module, used to perform welding defect detection on the first welding visualization image set and output welding visualization defect images.

[0017] The technical solution provided in this application has at least the following technical effects or advantages: by achieving the technical goal of deep integration and dynamic mapping of welding process and metal material property data, visualization image processing and BIM model, it can achieve the technical effect of real-time monitoring of welding defects, unified and standardized evaluation results, and intuitive presentation of welding quality in the three-dimensional model.

[0018] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in 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, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the BIM-based visual management method for metal welding in thermal power plants, as described in this application.

[0021] Figure 2This is a schematic diagram of the structure of the BIM-based visual management system for metal welding in thermal power plants, as described in this application.

[0022] Figure labeling: Module 1 for acquiring metal welding material information, Module 2 for determining the granularity of visualization pixel processing, Module 3 for obtaining the first set of welding visualization images, and Module 4 for outputting welding visualization defect images. Detailed Implementation

[0023] This application provides a BIM-based visual management method and system for metal welding in thermal power plants. It addresses the technical problems in existing technologies, such as the lack of a single welding quality monitoring method, the absence of dynamic fusion analysis of material properties and the welding process, and the lack of visualization integration with a BIM platform. These problems lead to untimely detection of welding defects, inconsistent and unintuitive inspection results, and ultimately affect the welding quality control level of key equipment and the overall safety and reliability of thermal power plant operations. The application achieves the technical goal of deep integration and dynamic mapping of welding process data, metal material property data, visual image processing, and BIM models, enabling real-time monitoring of welding defects, standardized and unified evaluation results, and intuitive presentation of welding quality in a 3D model.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a BIM-based visual management method for metal welding in thermal power plants, which is applied to a BIM-based visual management system for metal welding in thermal power plants. Specifically, it includes: S1: Obtain information on metal welding materials used for metal welding, and extract the metal material type, metal porosity, and welding surface roughness from the metal welding material information.

[0026] Specifically, before metal welding, information about the welding materials is collected. This information is closely related to the welding process, including the type of metal, its physical properties, and surface condition. The information includes the metal type, porosity, and surface roughness. The metal type refers to the specific material of the metal, such as steel, aluminum, or copper. Different types of metals require different welding processes and parameters due to differences in melting point, thermal conductivity, and strength. Porosity refers to the number and distribution of tiny pores within the metal, calculated as (area of ​​detected micropores or pores) / surface sampling area. It affects the density and strength of the weld. Higher porosity makes it easier for pores or cracks to form during welding, thus reducing the quality of the weld joint. Surface roughness refers to the smoothness of the metal surface to be welded, expressed numerically as the average difference in surface unevenness. Higher surface roughness indicates a less smooth surface, which may prevent the solder from fully wetting and filling the weld, leading to defects such as incomplete fusion or slag inclusions.

[0027] S2: Establish pixel granularity mapping rules, input the metal material type, metal material porosity and welding surface roughness into the pixel granularity mapping rules, and determine the visual pixel processing granularity.

[0028] Furthermore, this application also includes: constructing a metal material sample library, including average porosity samples of metal materials under different metal material types and average surface roughness samples of metals; collecting welding defect image samples from the metal material sample library under different metal welding conditions, scoring the defect display quality of the welding defect image samples at different pixel granularities, and obtaining a pixel granularity solution set of the welding defect image samples that meets a preset defect display quality threshold; constructing a defect-granularity association knowledge base based on the welding defect image samples and the pixel granularity solution set; assigning weight coefficients to the different metal material types, average porosity samples of metal materials, and average surface roughness samples of metals to construct a first weight coefficient; and learning and updating the defect-granularity association knowledge base with the first weight coefficient to obtain pixel granularity mapping rules.

[0029] Furthermore, this application also includes: initializing a random forest regression model, including the number of random trees, the depth of the random trees, and the number of node levels of the random trees; encoding the different metal material types, the average porosity samples of the metal materials, and the average roughness samples of the metal surfaces as input to initialize the training set of the random forest regression model, and training it with the pixel granularity solution set as the supervision label to obtain the trained random forest regression model; extracting the learning weight coefficients of the trained random forest regression model; comparing the learning weight coefficients with the first weight coefficients, and updating the model based on the weight coefficient deviation when the accuracy of the trained random forest regression model is less than a preset threshold to obtain a converged random forest regression model; and outputting pixel granularity mapping rules based on the converged random forest regression model.

[0030] Furthermore, this application also includes: determining whether the materials used for metal welding are the same metal material; if the materials used for metal welding are the same metal material, determining whether a transition metal material is used in the metal welding process; if a transition metal material is used, extracting the transition metal material type, transition metal material porosity, and transition metal surface roughness of the transition metal material; inputting the transition metal material type, transition metal material porosity, and transition metal surface roughness into the pixel granularity mapping rule to determine the transition visualization pixel processing granularity; and weightedly fusing the transition visualization pixel processing granularity and the visualization pixel processing granularity to output the fused visualization pixel processing granularity.

[0031] Furthermore, this application also includes: if the materials used for metal welding are different metal materials, extracting the first metal welding material information corresponding to the first end welding metal and the second metal welding material information corresponding to the second end welding metal; the pixel granularity mapping rule obtains the first visual pixel processing granularity and the second visual pixel processing granularity based on the first metal welding material information and the second metal welding material information respectively; and weightedly fusing the first visual pixel processing granularity and the second visual pixel processing granularity to output the fused visual pixel processing granularity.

[0032] Furthermore, this application also includes: if the materials used for metal welding are different metal materials, and the metal welding process does not use a transition metal material, the first visual pixel processing granularity and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity; if the materials used for metal welding are different metal materials, and the metal welding process uses a transition metal material, the transition visual pixel processing granularity, the first visual pixel processing granularity, and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity.

[0033] Specifically, constructing a metal material sample library refers to collecting different types of metal materials and recording statistical samples of their physical properties. The metal material sample library includes average porosity samples and average surface roughness samples for different metal material types. The average porosity sample describes the ratio of the internal area (detected micropores or pores) to the surface sampling area, reflecting the material's density; the average surface roughness sample describes the average value of the surface's height fluctuations, reflecting its smoothness.

[0034] Next, image samples of welding defects under different metal welding conditions are collected from a metal material sample library, such as weld porosity, cracks, or incomplete penetration defects. The welding defect image samples are then scored for display quality at different pixel granularities to measure whether the defects are clearly visible in the image. When the clarity of a welding defect image sample at different pixel granularities exceeds a preset defect display quality threshold, the corresponding pixel granularity of the welding defect image sample is selected into the pixel granularity solution set.

[0035] Then, based on the welding defect image samples and the corresponding pixel granularity solution set, a knowledge base relating defects to granularity is constructed. This knowledge base stores a dataset of defect types and optimal pixel granularity matching relationships, helping to quickly determine the optimal visualization scheme for different defects under different conditions. For example, porosity defects tend to be clearly displayed at high granularity, while crack defects can be clearly identified at medium granularity.

[0036] Next, weight coefficients are assigned to different metal material types, average porosity samples of metal materials, and average surface roughness samples of metals, assigning different levels of importance to different features and constructing the first weight coefficient. For example, in high-precision welding, the average surface roughness sample of metals may be assigned a weight of 0.6, while the average porosity sample of metal materials may be assigned a weight of 0.3, and the metal material type may be assigned a weight of 0.1, which can highlight the impact of key features on the visualization of defects.

[0037] Initialize a random forest regression model to establish a prediction model based on ensemble learning. Here, the number of random trees represents the number of independent decision trees in the random forest; the depth of the random tree represents the hierarchical complexity of each tree in feature partitioning; and the node level of the random tree refers to the number of branches in the random tree that can be used to partition the data.

[0038] Next, samples of different metal material types, average porosity of metal materials, and average surface roughness of metals are encoded and converted into numerical vectors that the model can recognize. The encoded data serves as the input to form the training set for initializing the random forest regression model, while the pixel-level solution set serves as the supervision label. The supervision label is used to provide correct answers for the initial random forest regression model to learn, thereby obtaining the trained random forest regression model.

[0039] Then, the learning weight coefficients of the trained random forest regression model are extracted. The feature importance distribution obtained during the training process is used to reflect the degree of contribution of different input features to the final output. For example, the random forest regression model may determine that the weight of surface roughness is 0.5, the weight of porosity density is 0.3, and the weight of metal type is 0.2, reflecting the model's learning preferences.

[0040] Next, the learned weight coefficients are compared with the first weight coefficients. The first weight coefficients are manually set baseline assignments, while the learned weight coefficients are assignments obtained by the model through self-learning. If the accuracy of the trained random forest regression model is less than a preset threshold, the weight coefficients are updated based on the deviation between the two. By continuously adjusting the parameters of the random forest regression model, the training results tend to stabilize, ultimately resulting in a converged random forest regression model. The converged random forest regression model outputs pixel-level mapping rules, which can provide a basis for selecting the optimal granularity for the visualization of welding defects under different metal material conditions.

[0041] During metal welding, it is necessary to determine whether the two welding materials used are the same metal material. The same metal material means that the metal material type, the pore density of the metal material, and the surface roughness of the weld are completely consistent.

[0042] When different metals are used for metal welding, it is necessary to extract the welding material information for both ends of the weld separately, namely, the first metal welding material information corresponding to the welding metal at the first end and the second metal welding material information corresponding to the welding metal at the second end. For example, one end is steel and the other end is copper.

[0043] Subsequently, the pixel granularity mapping rule independently calculates different visualization pixel processing granularities based on the welding material information of each of the two metals. Visualization pixel processing granularity refers to an adjustable parameter for the degree of mesh subdivision and texture mapping refinement of the weld surface during BIM 3D rendering. The first visualization pixel processing granularity is derived from the welding material information of the first metal, while the second visualization pixel processing granularity is derived from the welding material information of the second metal. A larger granularity value indicates a finer image division, enabling the display of more minute defects.

[0044] Next, the two different visualization pixel processing granularities are weighted and fused. Weighted fusion is a comprehensive method that assigns a weighted proportion to the first and second visualization pixel processing granularities according to the degree of influence of each metal on the welding image. For example, if the weight of steel is set to 0.4 and the weight of copper is set to 0.6, the final fusion result will be biased towards the granularity requirements of copper, thus making the overall image more reflective of the potential defects of copper during the welding process. The output fused visualization pixel processing granularity then becomes the unified standard for subsequent welding image visualization and defect detection.

[0045] If the determination result indicates the same metal, further investigation is needed to determine whether a transition metal material was introduced during the welding process. A transition metal material is an auxiliary material added to the welding interface of different metals, or used in welding of the same metal to improve weld performance, thereby reducing differences in thermal expansion coefficients, increasing weld strength, or preventing crack formation. For example, titanium is used as a transition metal material in nickel alloy welding to improve the corrosion resistance of the joint. If a transition metal material is used, it is necessary to extract the type of transition metal, its porosity, and its surface roughness for subsequent mapping calculations.

[0046] Next, the transition metal material type, porosity, and surface roughness are input into the pixel granularity mapping rule. The pixel granularity mapping rule is an algorithmic model that determines the pixel processing level based on the relationship between metal properties and the fineness of the visualized image. The transition visualization pixel processing granularity is the image accuracy requirement calculated based on the transition metal material type, porosity, and surface roughness, which determines whether subsequent welding defects can be clearly presented in the image. For example, when the porosity of the transition metal is high, the mapping rule may require finer pixel division so that potential porosity defects can be magnified and displayed.

[0047] Finally, the transition visualization pixel processing granularity is merged with the original visualization pixel processing granularity using a weighted fusion method to obtain a fused visualization pixel processing granularity. Weighted fusion refers to allocating proportions according to the degree of influence of different weight parameters on welding imaging accuracy. For example, the weight of transition metal materials is 0.4, and the weight of base metal materials is 0.6. Then the final pixel granularity result is the weighted output of the two, which can simultaneously take into account the characteristics of base materials and transition materials, improving the scientificity and adaptability of welding visualization management.

[0048] When different metals are used for metal welding, and no transition metal is used in the welding process, the visualization pixel granularity obtained from both metals needs to be weighted and fused. Weighted fusion combines the two results into a unified value by allocating a ratio. For example, if the steel granularity at the first end is 12 units and the aluminum granularity at the second end is 18 units, and the weights are allocated to 0.5 and 0.5 respectively, the final output fused granularity will be 15 units, thus balancing the visualization requirements of the two different metals.

[0049] In another scenario, if the materials used for metal welding are different metals, and the welding process employs a transition metal, three granularities need to be considered simultaneously: the granularity of the transition visualization pixels, the granularity of the first visualization pixels, and the granularity of the second visualization pixels. During weighted fusion, the proportions can be allocated based on the actual welding contribution. For example, if the granularity of the first-end steel is 12 units, the granularity of the second-end copper is 20 units, and the granularity of the transition metal nickel is 16 units, and the weights are 0.3, 0.3, and 0.4 respectively, the final fusion result will be approximately 16 units, ensuring that the role of the transition metal in adjusting the welding visualization accuracy is fully reflected.

[0050] S3: Collect an initial set of welding visualization images of the metal welding process, connect to the BIM visualization rendering engine platform, and perform pixel grid granular processing on the initial set of welding visualization images according to the visualization pixel processing granularity to obtain the first set of welding visualization images.

[0051] Specifically, during metal welding, an initial set of visual images of the welding process is acquired in real time using camera equipment or sensors, recording the arc state, molten pool morphology, and weld surface condition. These raw images form the basis for subsequent analysis and processing, providing a comprehensive view of the entire welding process.

[0052] Subsequently, the BIM visualization rendering engine platform is a tool that combines images with BIM 3D models. Connecting the initial set of welding visualization images to the BIM visualization rendering engine platform allows the images of the welding process to be mapped to the structural model of the power plant, facilitating subsequent analysis and management.

[0053] Next, the initial welding visualization image set is processed according to the obtained visualization pixel processing granularity, decomposing it into grid units of different sizes, each unit consisting of several pixels. The smaller the pixel grid granularity, the denser the grid, the more refined the image display, and the more minute defects can be captured. Finally, the processed images form the first welding visualization image set, containing granularized image data that more clearly presents the weld condition, facilitating subsequent defect detection and analysis.

[0054] S4: Perform welding defect detection on the first set of welding visualization images and output welding visualization defect images.

[0055] Furthermore, this application also includes: extracting feature values ​​of each welding visualization image in the first welding visualization image set with respect to grid pixel blocks, the feature values ​​including average gray value, average color value, gradient contrast and standard deviation; calculating the defect probability of each grid pixel block based on the feature values, marking grid pixel blocks with defect probabilities greater than preset defect probabilities, and performing color rendering and marking on the BIM visualization rendering engine platform to output welding visualization defect images.

[0056] Specifically, feature values ​​of each welding visualization image in the first welding visualization image set are extracted with respect to grid pixel blocks. This involves dividing the entire image into multiple small grid regions, each called a grid pixel block, and then calculating the feature values ​​of these grid pixel blocks. The feature values ​​include the average grayscale value, the average color value, the gradient contrast, and the standard deviation. The average grayscale value measures the brightness of the grid pixel block; the average color value reflects the overall color balance within the grid pixel block; the gradient contrast indicates the magnitude of pixel intensity variation within the grid pixel block; and the standard deviation is an indicator describing the dispersion of pixel distribution within the grid pixel block, used to reveal whether there are abnormal fluctuations in the grid pixel block.

[0057] Next, the defect probability of each grid pixel block is calculated based on the feature values. That is, by inputting the feature values ​​into a statistical model or classification algorithm, the likelihood of a defect in the grid pixel block is determined. Grid pixel blocks with defect probabilities greater than a preset defect probability are marked, thus providing a more intuitive basis for judgment.

[0058] Then, color rendering and marking are performed on the BIM visualization rendering engine platform. Pixels identified as defects are highlighted with different colors to visually distinguish defective areas from normal areas. For example, defective areas are rendered in red, while non-defective areas retain their original color. Finally, a weld visualization defect image is output, which is an analyzed and rendered image in which defects are clearly marked for easy inspection by engineers.

[0059] Furthermore, this application also includes: establishing a connection with the BIM visualization rendering engine platform through an application programming interface (API) and issuing visualization pixel processing commands; the BIM visualization rendering engine platform performing coordinate transformation on the acquired initial welding visualization image set in the BIM model coordinate system based on the visualization pixel processing commands to obtain a BIM welding visualization mapping image set; and performing pixel grid granular processing on the BIM welding visualization mapping image set according to the visualization pixel processing granularity to obtain a first welding visualization image set.

[0060] Specifically, a connection is established with the BIM visualization rendering engine platform through an Application Programming Interface (API), enabling external programs to interact with the BIM engine for data exchange. An API is a predefined communication method that allows different software systems to collaborate smoothly; the BIM visualization rendering engine platform is a tool for 3D modeling and simulation of architecture, machinery, or welding, capable of 3D visualization of welding scenes. Issuing visualization pixel processing commands, i.e., sending processing instructions to the BIM platform through the API, is used to control pixel-level image processing.

[0061] Next, the BIM visualization rendering engine platform performs coordinate transformation on the initial set of welding visualization images in the BIM model coordinate system based on visualization pixel processing commands. This involves mapping the pixel positions in the two-dimensional images to the spatial coordinates of the three-dimensional BIM model. The BIM model coordinate system is a three-dimensional spatial positioning system used to determine the specific location of welded components in the virtual environment. Coordinate transformation refers to mapping the pixel information of the original initial set of welding visualization images to their actual geometric positions in the BIM model, thereby ensuring the alignment of the image data with the three-dimensional model and obtaining a BIM welding visualization mapped image set.

[0062] Then, the BIM welding visualization image set is processed at the pixel grid granularity according to the visualization pixel processing granularity, and the images are re-divided and analyzed according to the set pixel granularity standard. For example, if the visualization pixel processing granularity is set to 10, it means that every 10 pixels are a grid block to extract features, thereby obtaining the first welding visualization image set, which can provide a basis for subsequent defect identification and quality analysis.

[0063] Furthermore, this application also includes: the BIM visualization rendering engine platform divides the BIM welding visualization mapping image set into a grid according to the parameterized grid components at the visualization pixel processing granularity, and outputs a first welding visualization image set; wherein, each welding visualization image in the first welding visualization image set includes a set of grid pixel blocks corresponding to the pixel grid granularity, and the set of grid pixel blocks is stored according to the index position.

[0064] Specifically, the BIM visualization rendering engine platform divides the BIM welding visualization mapping image set into a mesh based on parametric mesh components and a visual pixel processing granularity. This means using parametric modeling methods, the welding images are subdivided according to the components of the 3D model and a set pixel granularity. Parametric mesh components refer to 3D units whose size, shape, and position can be adjusted via parameters. The visual pixel processing granularity determines the number of pixels covered by each mesh unit. For example, a visual pixel processing granularity of 10 can mean extracting features by dividing the image set into 10×10 pixels or 10 pixels per mesh block. This structured BIM welding visualization mapping image set provides a foundation for analysis and visualization, ultimately outputting the first welding visualization image set—the processed image set.

[0065] Next, each welding visualization image in the first welding visualization image set includes a set of grid pixel blocks corresponding to a pixel grid granularity. This means that each image in the first welding visualization image set is divided into multiple grids, and each grid block contains a number of pixels. The set of grid pixel blocks is stored according to index position, that is, each grid pixel block has a unique number or coordinate index for fast access and subsequent processing. For example, if a welding visualization image is divided into a 20×20 grid, there are a total of 400 grid pixel blocks, and the index of each block can range from 1 to 400, thus enabling precise location when extracting features or calculating defect probabilities.

[0066] In summary, the BIM-based visual management method for metal welding in thermal power plants provided in this application has the following technical effects: by achieving the technical goal of deep integration and dynamic mapping of welding process and metal material property data, visual image processing and BIM model, it can achieve the technical effects of real-time monitoring of welding defects, unified and standardized evaluation results, and intuitive presentation of welding quality in the three-dimensional model.

[0067] Example 2: Based on the same inventive concept as the BIM-based visual management method for metal welding in thermal power plants described in the previous examples, this application also provides a BIM-based visual management system for metal welding in thermal power plants. Please refer to the appendix. Figure 2The system includes: a metal welding material information acquisition module 1, used to acquire metal welding material information for metal welding, and extract the metal material type, metal porosity, and welding surface roughness of the metal welding material information; a visualization pixel processing granularity determination module 2, used to establish pixel granularity mapping rules, inputting the metal material type, metal porosity, and welding surface roughness into the pixel granularity mapping rules to determine the visualization pixel processing granularity; a first welding visualization image set acquisition module 3, used to acquire an initial welding visualization image set of the metal welding process, connect to the BIM visualization rendering engine platform, and perform pixel grid granularity processing on the initial welding visualization image set according to the visualization pixel processing granularity to obtain a first welding visualization image set; and a welding visualization defect image output module 4, used to perform welding defect detection on the first welding visualization image set and output welding visualization defect images.

[0068] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used to: determine whether the materials used for metal welding are the same metal material; if the materials used for metal welding are the same metal material, determine whether a transition metal material is used in the metal welding process; if a transition metal material is used, extract the transition metal material type, transition metal porosity, and transition metal surface roughness of the transition metal material; input the transition metal material type, transition metal porosity, and transition metal surface roughness into the pixel granularity mapping rule to determine the transition visualization pixel processing granularity; and weightedly fuse the transition visualization pixel processing granularity and the visualization pixel processing granularity to output the fused visualization pixel processing granularity.

[0069] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used for: if the materials used for metal welding are different metal materials, extracting the first metal welding material information corresponding to the first end welding metal and the second metal welding material information corresponding to the second end welding metal; the pixel granularity mapping rule obtains the first visualization pixel processing granularity and the second visualization pixel processing granularity according to the first metal welding material information and the second metal welding material information respectively; and weightedly fusing the first visualization pixel processing granularity and the second visualization pixel processing granularity to output the fused visualization pixel processing granularity.

[0070] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used to: if the materials used for metal welding are different metal materials and the metal welding process does not use transitional metal materials, the first visualization pixel processing granularity and the second visualization pixel processing granularity are weighted and fused to output the fused visualization pixel processing granularity; if the materials used for metal welding are different metal materials and the metal welding process uses transitional metal materials, the transitional visualization pixel processing granularity, the first visualization pixel processing granularity, and the second visualization pixel processing granularity are weighted and fused to output the fused visualization pixel processing granularity.

[0071] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used for: constructing a metal material sample library, including average porosity samples of metal materials under different metal material types and average surface roughness samples of metal materials; collecting welding defect image samples from the metal material sample library under different metal welding conditions, scoring the defect display quality of the welding defect image samples at different pixel granularities, and obtaining a pixel granularity solution set of the welding defect image samples that meets a preset defect display quality threshold; constructing a defect-granularity association knowledge base based on the welding defect image samples and the pixel granularity solution set; assigning weight coefficients to the different metal material types, average porosity samples of metal materials, and average surface roughness samples of metal materials to construct a first weight coefficient; and learning and updating the defect-granularity association knowledge base with the first weight coefficient to obtain pixel granularity mapping rules.

[0072] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used for: initializing a random forest regression model, including the number of random trees, the depth of the random trees, and the number of node levels of the random trees; encoding the samples of different metal material types, average porosity of metal materials, and average roughness of metal surfaces as input to initialize the training set of the random forest regression model, and training with the pixel granularity solution set as supervision labels to obtain the trained random forest regression model; extracting the learning weight coefficients of the trained random forest regression model; comparing the learning weight coefficients with the first weight coefficients, and updating the model based on the weight coefficient deviation when the accuracy of the trained random forest regression model is less than a preset threshold to obtain a converged random forest regression model; and outputting pixel granularity mapping rules based on the converged random forest regression model.

[0073] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used for: establishing a connection with the BIM visualization rendering engine platform through an application programming interface (API) and issuing visualization pixel processing commands; the BIM visualization rendering engine platform transforms the coordinates of the initial welding visualization image set obtained in the BIM model coordinate system based on the visualization pixel processing commands to obtain a BIM welding visualization mapping image set; and performing pixel grid granular processing on the BIM welding visualization mapping image set according to the visualization pixel processing granularity to obtain a first welding visualization image set.

[0074] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used for: the BIM visualization rendering engine platform dividing the BIM welding visualization mapping image set into a grid according to the parametric grid components at the visualization pixel processing granularity, and outputting a first welding visualization image set; wherein, each welding visualization image in the first welding visualization image set includes a set of grid pixel blocks corresponding to the pixel grid granularity, and the set of grid pixel blocks is stored according to the index position.

[0075] Furthermore, the BIM-based metal welding visualization management system for thermal power plants is also used to: extract feature values ​​of each welding visualization image in the first welding visualization image set with respect to grid pixel blocks, the feature values ​​including average gray value, average color value, gradient contrast and standard deviation; calculate the defect probability of each grid pixel block according to the feature values, mark grid pixel blocks with defect probabilities greater than preset defect probabilities, and perform color rendering and marking on the BIM visualization rendering engine platform to output welding visualization defect images.

[0076] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The BIM-based metal welding visualization management method and specific examples in Embodiment 1 described above are also applicable to the BIM-based metal welding visualization management system for thermal power plants in this embodiment. Through the foregoing detailed description of the BIM-based metal welding visualization management method for thermal power plants, those skilled in the art can clearly understand the BIM-based metal welding visualization management system for thermal power plants in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0077] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0078] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A BIM-based visual management method for metal welding in thermal power plants, characterized in that, The method includes: Obtain information on metal welding materials used for metal welding, and extract the metal material type, metal porosity, and welding surface roughness from the metal welding material information; Establish pixel granularity mapping rules, input the metal material type, metal porosity, and welding surface roughness into the pixel granularity mapping rules, and determine the visual pixel processing granularity; An initial set of welding visualization images of the metal welding process is collected, connected to a BIM visualization rendering engine platform, and the initial set of welding visualization images is processed at the pixel grid granularity according to the visualization pixel processing granularity to obtain a first set of welding visualization images. Perform welding defect detection on the first set of welding visualization images and output a welding defect visualization image; Methods for establishing pixel-level mapping rules include: Construct a metal material sample library, including average porosity samples of metal materials under different metal material types and average surface roughness samples of metals; Collect welding defect image samples from the metal material sample library under different metal welding conditions, score the defect display quality of the welding defect image samples at different pixel granularities, and obtain the pixel granularity solution set of the welding defect image samples that meets the preset defect display quality threshold. Based on the welding defect image samples and the pixel granularity solution set, a defect-granularity association knowledge base is constructed. Weighting coefficients are assigned to the different metal material types, the average porosity samples of the metal materials, and the average roughness samples of the metal surfaces to construct a first weighting coefficient. The pixel granularity mapping rule is obtained by learning and updating the defect-granularity association knowledge base using the first weight coefficient. The pixel granularity mapping rule is obtained by learning and updating the defect-granularity association knowledge base using the first weight coefficient. The method includes: Initialize the random forest regression model, including the number of random trees, the depth of the random trees, and the number of node levels in the random trees; The different metal material types, the average porosity samples of the metal materials, and the average roughness samples of the metal surfaces are encoded and used as input to initialize the training set of the random forest regression model. The pixel-level solution set is used as the supervision label for training to obtain the trained random forest regression model. Extract the learning weight coefficients of the trained random forest regression model; The learned weight coefficients are compared with the first weight coefficients. When the accuracy of the trained random forest regression model is less than a preset threshold, it is updated based on the weight coefficient deviation to obtain a converged random forest regression model. The pixel granularity mapping rule is output according to the converged random forest regression model.

2. The BIM-based visual management method for metal welding in thermal power plants as described in claim 1, characterized in that, Methods for determining the granularity of pixel processing for visualization also include: Determine whether the materials used for metal welding are the same metal material; If the materials used for metal welding are the same metal material, determine whether the metal welding process uses a transition metal material. If a transition metal material is used, extract the transition metal material type, transition metal material porosity, and transition metal surface roughness of the transition metal material. The transition metal material type, transition metal porosity, and transition metal surface roughness are input into the pixel granularity mapping rule to determine the transition visualization pixel processing granularity. The weighted fusion of the transition visualization pixel processing granularity and the visualization pixel processing granularity outputs the fused visualization pixel processing granularity.

3. The BIM-based visual management method for metal welding in thermal power plants as described in claim 2, characterized in that, Methods for determining whether the materials used for metal welding are the same metal also include: If the materials used for metal welding are different metal materials, extract the first metal welding material information corresponding to the first end welding metal and the second metal welding material information corresponding to the second end welding metal; The pixel granularity mapping rule obtains the first visual pixel processing granularity and the second visual pixel processing granularity based on the first metal welding material information and the second metal welding material information, respectively. The weighted fusion of the first and second visual pixel processing granularities outputs the fused visual pixel processing granularity.

4. The BIM-based visual management method for metal welding in thermal power plants as described in claim 3, characterized in that, If the materials used for metal welding are different metal materials, and the metal welding process does not use transition metal materials, the first visual pixel processing granularity and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity. If the materials used for metal welding are different metal materials, and the metal welding process uses a transitional metal material, the transitional visual pixel processing granularity, the first visual pixel processing granularity, and the second visual pixel processing granularity are weighted and fused to output the fused visual pixel processing granularity.

5. The BIM-based visual management method for metal welding in thermal power plants as described in claim 1, characterized in that, Establish a connection with the BIM visualization rendering engine platform through the application programming interface and issue visualization pixel processing commands; The BIM visualization rendering engine platform transforms the initial welding visualization image set in the BIM model coordinate system based on the visualization pixel processing command to obtain the BIM welding visualization mapping image set. The BIM welding visualization mapping image set is processed at the pixel grid granularity according to the visualization pixel processing granularity to obtain the first welding visualization image set.

6. The BIM-based visual management method for metal welding in thermal power plants as described in claim 5, characterized in that, The BIM visualization rendering engine platform divides the BIM welding visualization mapping image set into a mesh according to the parametric mesh components and the visualization pixel processing granularity, and outputs the first welding visualization image set. Each welding visualization image in the first welding visualization image set includes a set of grid pixel blocks corresponding to the pixel grid granularity, and the set of grid pixel blocks is stored according to the index position.

7. The BIM-based visual management method for metal welding in thermal power plants as described in claim 1, characterized in that, The method includes performing welding defect detection on the first set of welding visualization images and outputting welding defect visualization images. Extract the feature values ​​of each welding visualization image in the first welding visualization image set with respect to the grid pixel block. The feature values ​​include the average gray value, the average color value, the gradient contrast, and the standard deviation. The defect probability of each grid pixel block is calculated based on the feature value. Grid pixel blocks with defect probabilities greater than the preset defect probability are marked, and then color-rendered and marked on the BIM visualization rendering engine platform to output a visual image of welding defects.

8. A BIM-based visual management system for metal welding in thermal power plants, characterized in that, The steps for implementing the BIM-based visual management method for metal welding in thermal power plants according to any one of claims 1 to 7 include: The metal welding material information acquisition module is used to acquire metal welding material information for metal welding, and to extract the metal material type, metal porosity, and welding surface roughness of the metal welding material information. The visualization pixel processing granularity determination module is used to establish pixel granularity mapping rules. The metal material type, metal porosity, and welding surface roughness are input into the pixel granularity mapping rules to determine the visualization pixel processing granularity. The first welding visualization image set acquisition module is used to acquire an initial welding visualization image set of the metal welding process, connect to the BIM visualization rendering engine platform, and perform pixel grid granular processing on the initial welding visualization image set according to the visualization pixel processing granularity to obtain the first welding visualization image set. The welding visualization defect image output module is used to detect welding defects in the first set of welding visualization images and output welding visualization defect images.

Images