Complex structural part local corrosion analysis method and system
By using optical 3D scanning and supervised learning models, the problem of quantifying localized corrosion in large-mass complex structural components was solved, enabling precise assessment of the non-uniform corrosion type and development stage of complex structural components, thus overcoming the limitations of traditional methods.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-30
- Publication Date
- 2026-03-31
AI Technical Summary
Existing technologies cannot effectively quantify localized corrosion in large-mass, complex structural components, especially near curved surface interfaces. Traditional methods fail due to size effects, geometric limitations, and electrochemical interference.
Three-dimensional surface data of complex structural components are obtained by optical 3D scanning, and the data is fitted with the substrate geometry to extract corrosion morphology topological feature vectors. A supervised learning model is then used to classify corrosion types and development stages, and a local corrosion assessment report is generated.
It enables automatic identification of non-uniform corrosion types and precise determination of corrosion development stages for large-mass, complex structural components, breaking through the size and geometric limitations of traditional methods and providing a comprehensive and objective corrosion assessment.
Smart Images
Figure CN121767696A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of material corrosion testing technology, and in particular to a method and system for analyzing localized corrosion of complex structural components. Background Technology
[0002] Current analytical methods for localized corrosion of metal structural components mainly fall into three categories, but all have significant limitations: **Weight loss / weight gain methods (e.g., ASTM G31 standard):** These methods calculate the corrosion rate by measuring the mass change of the sample before and after corrosion. This method is only suitable for small, homogeneous material samples (typically <1 kg, <10 cm in size). Complex structural components (such as bridge joints and offshore platform support structures) are too large and irregularly shaped to be directly weighed or disassembled, rendering this method ineffective. **Electrochemical methods (e.g., polarization curves, impedance spectroscopy):** These methods require electrodes to be installed in the test area to monitor electrochemical signals. However, for complex structural components near interfaces (e.g., metal-cement contact areas), the limited geometric space makes electrode placement difficult. Furthermore, electrochemical methods reflect localized electrochemical activity and cannot directly quantify macroscopic corrosion volume loss, especially for assessing non-uniform corrosion (e.g., pitting corrosion, crevice corrosion). Surface observation methods (such as SEM and metallographic analysis) involve observing corrosion morphology under a microscope, but can only provide local microscopic information (micrometer scale) and cannot reflect macroscopic corrosion volume changes; in addition, it requires destructive sampling of the sample, which is not suitable for large structural components in service.
[0003] Existing technologies cannot meet the need for accurate quantitative analysis of localized corrosion in real-world environments (outdoor corrosion) of large-mass, complex structural components (such as metal-cement composite systems), especially when corrosion is concentrated near curved interface surfaces. Traditional methods fail due to size effects, geometric limitations, and electrochemical interference. Summary of the Invention
[0004] In view of the aforementioned problems, this application is hereby filed.
[0005] Therefore, this application provides a method and system for analyzing localized corrosion of complex structural components, which can solve the problems mentioned in the background art.
[0006] To solve the above-mentioned technical problems, this application provides the following technical solution: In a first aspect, this application provides a method for analyzing localized corrosion of complex structural components, comprising: acquiring original three-dimensional surface data of the target area of the complex structural component after exposure to a corrosive environment, and extracting the matrix geometry type identifier and corrosion morphology topological feature vector from the original three-dimensional surface data; The corrosion morphology topological feature vector is input into the main classifier to obtain the main category classification result. The main category classification result is the local corrosion type label, which includes pitting corrosion, crevice corrosion, or trench corrosion. Based on the combination of the main category classification result and the matrix geometry type identifier, a target sub-classifier is determined from a preset set of sub-classifiers, wherein the set of sub-classifiers includes a first sub-classifier corresponding to the combination of spherical surface and pitting corrosion, a second sub-classifier corresponding to the combination of cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to the combination of planar surface and trench corrosion. The corrosion morphology topological feature vector is input into the target sub-classifier to obtain the sub-class classification result, which represents the corrosion development stage under the same local corrosion type. Based on the main category classification results, the sub-category classification results, and the corrosion volume loss in the corrosion morphology topological feature vector, a local corrosion assessment report for the structural component is generated.
[0007] Preferably, the extraction of the matrix geometry type identifier and corrosion morphology topological feature vector includes: Optical 3D scanning was performed on the target area of a complex structural component after exposure to a corrosive environment to obtain the original 3D surface data. Based on the structural characteristics of the complex structural component, the base geometry type corresponding to the target area is determined, and a base geometry type identifier is generated. The base geometry type identifier is one of a sphere, a cylinder, or a plane. The original three-dimensional surface data is preprocessed to remove scanning noise and invalid point clouds, resulting in valid three-dimensional surface data. The corrosion morphology topological feature vector is extracted from the effective three-dimensional surface data. The corrosion morphology topological feature vector includes the average depth, the arithmetic mean deviation of the profile Ra, the slope of the corrosion pit edge, and the corrosion volume loss.
[0008] Preferably, obtaining the main category classification result includes: Construct a master classifier, which is a supervised learning model. The samples used to train the master classifier include residual surface data and corresponding local corrosion type labels. The corrosion morphology topological feature vector is normalized so that the numerical range of each dimension of the corrosion morphology topological feature vector is unified to the interval between 0 and 1. The normalized corrosion morphology topological feature vector is input into the main classifier to calculate the predicted probability of the labels of the three local corrosion types: pitting corrosion, crevice corrosion, and trench corrosion. The label of the localized corrosion type with the highest predicted probability was selected as the main category classification result.
[0009] Preferably, determining the target sub-classifier from a preset set of sub-classifiers includes: A set of sub-classifiers is established, which includes a first sub-classifier, a second sub-classifier, and a third sub-classifier. The first sub-classifier corresponds to the combination of spherical surface and pitting corrosion, the second sub-classifier corresponds to the combination of cylindrical surface and crevice corrosion, and the third sub-classifier corresponds to the combination of planar surface and trench corrosion. The main category classification result is combined with the base geometry type identifier to form a combined identifier; The combined identifier is matched with the combination corresponding to each subclassifier in the subclassifier set; When the combination identifier matches the combination corresponding to a certain sub-classifier, the sub-classifier is determined as the target sub-classifier.
[0010] Preferably, the sub-category classification results include: Construct a target sub-classifier, which is a supervised learning model. The samples used to train the target sub-classifier include residual surface data, corresponding local corrosion type labels, and corrosion development stage labels. The corrosion morphology topological feature vector is input into the target sub-classifier to calculate the predicted probability of each candidate corrosion development stage. The corrosion development stage with the highest predicted probability is selected as the sub-category classification result, which represents the corrosion development stage under the same local corrosion type.
[0011] Preferably, the generated structural component localized corrosion assessment report includes: The type of localized corrosion is determined based on the main category classification results; The corrosion development stage corresponding to the localized corrosion type is determined based on the sub-category classification results; Extract corrosion volume loss from the corrosion morphology topological feature vector; The localized corrosion type, the corrosion development stage, and the corrosion volume loss are combined to generate a localized corrosion assessment report for the structural component.
[0012] Preferably, before extracting the topological feature vector of the erosion morphology, the method further includes: Based on the matrix geometry type identifier, the corresponding surface mathematical model is selected to fit the original three-dimensional surface data to generate an ideal matrix surface. When the matrix geometry type identifier is a sphere, the least squares method is used to fit the sphere equation; when the matrix geometry type identifier is a cylinder, the least squares method is used to fit the cylinder equation. The original three-dimensional surface data and the ideal substrate surface are subtracted point by point to obtain the residual surface data; The residual surface data is segmented by depth threshold, and regions with a depth greater than the corrosion threshold are extracted as effective corrosion regions. The corrosion morphology topological feature vector is extracted based on the residual surface data of the effective corrosion region.
[0013] Secondly, this application also provides a system for analyzing localized corrosion of complex structural components, including: a data acquisition module, used to acquire the original three-dimensional surface data of the target area of the complex structural component after exposure to a corrosive environment, and to extract the matrix geometry type identifier and corrosion morphology topological feature vector from the original three-dimensional surface data; The main classification module is used to input the corrosion morphology topological feature vector into the main classifier to obtain the main category classification result. The main category classification result is a local corrosion type label, which includes pitting corrosion, crevice corrosion, or trench corrosion. The subclassification selection module is used to determine a target subclassifier from a preset subclassifier set based on the combination of the main category classification result and the matrix geometry type identifier. The subclassifier set includes a first subclassifier corresponding to the combination of spherical surface and pitting corrosion, a second subclassifier corresponding to the combination of cylindrical surface and crevice corrosion, and a third subclassifier corresponding to the combination of planar surface and trench corrosion. The development stage discrimination module is used to input the corrosion morphology topological feature vector into the target sub-classifier to obtain the sub-class classification result, which represents the corrosion development stage under the same local corrosion type. The assessment report generation module is used to generate a local corrosion assessment report for structural components based on the main category classification results, the sub-category classification results, and the corrosion volume loss in the corrosion morphology topological feature vector.
[0014] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: Acquire the original three-dimensional surface data of the target area of a complex structural component after exposure to a corrosive environment, and extract the matrix geometry type identifier and corrosion morphology topological feature vector from the original three-dimensional surface data; The corrosion morphology topological feature vector is input into the main classifier to obtain the main category classification result. The main category classification result is the local corrosion type label, which includes pitting corrosion, crevice corrosion, or trench corrosion. Based on the combination of the main category classification result and the matrix geometry type identifier, a target sub-classifier is determined from a preset set of sub-classifiers, wherein the set of sub-classifiers includes a first sub-classifier corresponding to the combination of spherical surface and pitting corrosion, a second sub-classifier corresponding to the combination of cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to the combination of planar surface and trench corrosion. The corrosion morphology topological feature vector is input into the target sub-classifier to obtain the sub-class classification result, which represents the corrosion development stage under the same local corrosion type. Based on the main category classification results, the sub-category classification results, and the corrosion volume loss in the corrosion morphology topological feature vector, a local corrosion assessment report for the structural component is generated.
[0015] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps: Acquire the original three-dimensional surface data of the target area of a complex structural component after exposure to a corrosive environment, and extract the matrix geometry type identifier and corrosion morphology topological feature vector from the original three-dimensional surface data; The corrosion morphology topological feature vector is input into the main classifier to obtain the main category classification result. The main category classification result is the local corrosion type label, which includes pitting corrosion, crevice corrosion, or trench corrosion. Based on the combination of the main category classification result and the matrix geometry type identifier, a target sub-classifier is determined from a preset set of sub-classifiers, wherein the set of sub-classifiers includes a first sub-classifier corresponding to the combination of spherical surface and pitting corrosion, a second sub-classifier corresponding to the combination of cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to the combination of planar surface and trench corrosion. The corrosion morphology topological feature vector is input into the target sub-classifier to obtain the sub-class classification result, which represents the corrosion development stage under the same local corrosion type. Based on the main category classification results, the sub-category classification results, and the corrosion volume loss in the corrosion morphology topological feature vector, a local corrosion assessment report for the structural component is generated.
[0016] Implementing this application has the following beneficial effects: This invention provides a method for analyzing localized corrosion of large-mass, non-removable complex structural components. It acquires high-resolution morphological data of the target area through optical three-dimensional scanning and combines this with surface fitting and background subtraction based on the substrate geometry (e.g., spherical or cylindrical surfaces), effectively eliminating the interference of structural curvature on corrosion morphology extraction and significantly improving the accuracy of corrosion volume calculation. Based on this, it extracts multi-dimensional topological features such as average depth, profile arithmetic mean deviation Ra, corrosion pit edge slope, and corrosion volume loss from residual surface data, achieving automatic identification of non-uniform corrosion types such as pitting corrosion, crevice corrosion, and trench corrosion, and precise discrimination of corrosion development stages. Finally, it integrates corrosion type, development stage, and volume loss to generate a localized corrosion assessment report for the structural component. This method overcomes the technical bottlenecks of traditional weight loss methods (limited by sample size), electrochemical methods (difficult to deploy electrodes), and surface observation methods (unable to quantify macroscopic corrosion). It combines large-size adaptability, surface geometry compatibility, three-dimensional corrosion quantification capability, and non-uniform corrosion characterization capability, providing comprehensive, objective, and feasible technical support for durability evaluation and maintenance decisions of in-service complex structural components. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is an overall flowchart of a method for analyzing localized corrosion of complex structural components, as described in this application. Figure 2 This is a computer device diagram of a method for analyzing localized corrosion of complex structural components, which is the subject of this application. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0020] In one exemplary embodiment, such as Figure 1 As shown, a method for analyzing localized corrosion in complex structural components is provided, including: S100: Acquire the original three-dimensional surface data of the target area of a complex structural component after exposure to a corrosive environment, and extract the matrix geometry type identifier and corrosion morphology topological feature vector from the original three-dimensional surface data; S101: Perform optical 3D scanning on the target area of a complex structural component after exposure to a corrosive environment to obtain the original 3D surface data; In an optional embodiment, optical 3D scanning is performed using a white light interferometer, and high-resolution imaging of micron-scale corrosion pits is achieved by adjusting the coherence length of the light source and the numerical aperture of the objective lens.
[0021] In an optional embodiment, the optical 3D scanning employs a laser confocal scanner to simultaneously record surface reflection intensity information during the scanning process, which is used to assist in subsequent noise point identification.
[0022] In an optional embodiment, for a large target area spanning a weld or metal-cement interface, the scanning area is divided into multiple sub-fields of view, local three-dimensional data is acquired sequentially, and then automatically stitched together into complete original three-dimensional surface data using a feature point matching algorithm.
[0023] In an optional embodiment, the target area is surface-cleaned before scanning to remove dust and loose corrosion products, but retains a firmly attached corrosion layer to ensure that the original three-dimensional surface data accurately reflects the corrosion morphology of the substrate.
[0024] S102: Based on the structural characteristics of the complex structural component, determine the base geometry type corresponding to the target area and generate a base geometry type identifier. The base geometry type identifier can be one of a sphere, a cylinder, or a plane. In an optional embodiment, structural features are obtained in advance through design drawings or CAD models, and the system automatically compares the coordinates of the target area with the geometric attributes of the model and outputs the corresponding base geometry type identifier.
[0025] In an optional embodiment, if no design drawings are available, the determination is made based on the global curvature distribution of the original three-dimensional surface data: when the principal curvatures are all close to non-zero constants and have the same sign, it is determined to be a sphere; when one principal curvature is zero and the other principal curvature is non-zero, it is determined to be a cylinder; when both principal curvatures are close to zero, it is determined to be a plane.
[0026] In an optional embodiment, based on embodiment 1, if the target area is located on the outer surface of the bridge steel support, the base geometry type is directly set to cylindrical; if it is located in the tank head weld area, it is set to spherical; if it is located in the flat plate connector, it is set to planar.
[0027] S103: Preprocess the original 3D surface data to remove scanning noise and invalid point clouds, and obtain valid 3D surface data; In an optional embodiment, a median filtering algorithm is used to denoise the original three-dimensional surface data, and the size of the filtering window is dynamically set to 3×3 to 5×5 pixels according to the scanning resolution.
[0028] In an optional embodiment, outliers are identified by setting a height abrupt change threshold (e.g., the height difference between adjacent points > 20 μm) and such points are marked as invalid point clouds and removed.
[0029] In an optional embodiment, after removing invalid point clouds, bilinear interpolation or radial basis function interpolation is used to fill in the missing regions to ensure that the valid 3D surface data is spatially continuous and free of voids.
[0030] S104: Extract corrosion morphology topological feature vectors from effective three-dimensional surface data. The corrosion morphology topological feature vectors include average depth, profile arithmetic mean deviation Ra, corrosion pit edge slope, and corrosion volume loss. In an optional embodiment, the average depth is obtained by calculating the average depth of all corrosion region points relative to the fitted substrate surface, wherein the corrosion regions are segmented by a depth threshold (>10 μm).
[0031] In an optional embodiment, the profile arithmetic mean deviation Ra is calculated according to ISO 4287, which is the arithmetic mean of the absolute values of the height deviations of all sampling points within the corrosion area.
[0032] In an optional embodiment, the slope of the corrosion pit edge is obtained by calculating the ratio of the elevation difference between the boundary point of the corrosion zone and the non-corrosion point in its neighborhood to the horizontal distance, in degrees or dimensionless slope.
[0033] In an optional embodiment, the corrosion volume loss is calculated using the voxel integration method: the corrosion region is divided into 1μm³ voxel units in three-dimensional space, the number of all voxels with a depth less than the substrate surface is counted and multiplied by the volume of a single voxel to obtain the total corrosion volume loss.
[0034] It should be noted that before extracting the topological feature vector of the erosion morphology, the following steps are also included: Based on the matrix geometry type identifier, the corresponding surface mathematical model is selected to fit the original three-dimensional surface data to generate an ideal matrix surface. When the matrix geometry type identifier is a sphere, the least squares method is used to fit the sphere equation; when the matrix geometry type identifier is a cylinder, the least squares method is used to fit the cylinder equation. The residual surface data is obtained by point-by-point subtraction between the original three-dimensional surface data and the ideal substrate surface; Depth threshold segmentation is performed on the residual surface data, and regions with depths greater than the corrosion threshold are extracted as effective corrosion regions; The corrosion morphology topological feature vector is extracted based on the residual surface data of the effective corrosion region.
[0035] S200: Input the corrosion morphology topological feature vector into the main classifier to obtain the main category classification result. The main category classification result is the local corrosion type label, which includes pitting corrosion, crevice corrosion, or trench corrosion. S201: Construct the master classifier. The master classifier is a supervised learning model. The samples used to train the master classifier include residual surface data and corresponding local corrosion type labels. In an optional embodiment, the supervised learning model adopts a random forest model. The samples used to train the random forest model are obtained as follows: multiple metal-cement composite samples are placed in a salt spray test chamber for accelerated corrosion. After corrosion, the target area of each sample is optically scanned in three dimensions and surface fitting and background subtraction are performed to obtain residual surface data. Three technicians independently label the local corrosion type based on the morphological characteristics of the residual surface data. When the labeling results of at least two technicians are consistent, the labeling result is used as the final local corrosion type label and forms a training sample with the corresponding residual surface data.
[0036] In an optional embodiment, the localized corrosion type label is defined as pitting corrosion, crevice corrosion, or trench corrosion, wherein the residual surface data corresponding to pitting corrosion is represented as isolated circular or near-circular pits, the residual surface data corresponding to crevice corrosion is represented as a narrow corrosion area extending along the structural crevice, and the residual surface data corresponding to trench corrosion is represented as linear corrosion grooves continuously distributed along a certain direction.
[0037] In an optional embodiment, when training the main classifier, each residual surface data is converted into a corrosion morphology topological feature vector, which includes the average depth, the arithmetic mean deviation of the profile Ra, the slope of the corrosion pit edge, and the corrosion volume loss. The corrosion morphology topological feature vector is used as the input feature, and the corresponding local corrosion type label is used as the output label to complete the training of the main classifier.
[0038] S202: Normalize the corrosion morphology topological feature vector to unify the numerical range of each dimension in the corrosion morphology topological feature vector to the interval of 0 to 1. In an optional embodiment, the normalization process employs a minimum-maximum scaling method to process the average depth, profile arithmetic mean deviation Ra, corrosion pit edge slope, and corrosion volume loss in the corrosion morphology topological feature vector. The original value range of the average depth is set to 0 micrometers to 100 micrometers, the original value range of the profile arithmetic mean deviation Ra is set to 0 micrometers to 10 micrometers, the original value range of the corrosion pit edge slope is set to 0 degrees to 90 degrees, and the original value range of the corrosion volume loss is set to 0 cubic millimeters to 10 cubic millimeters.
[0039] In an optional embodiment, the original value range is determined based on the parameter extreme values observed in the accelerated corrosion experiment and written into the parameter configuration file during system initialization. During normalization, the value range in the parameter configuration file is directly called for calculation.
[0040] S203: Input the normalized corrosion morphology topological feature vector into the main classifier and calculate the predicted probability of the labels of the three local corrosion types: pitting corrosion, crevice corrosion and trench corrosion. In an optional embodiment, the main classifier is a random forest model containing multiple decision trees. Each decision tree performs node splitting based on the feature values in the normalized corrosion morphology topological feature vector. Finally, each decision tree outputs a local corrosion type label. The local corrosion type labels output by all decision trees are counted, and the occurrence frequency of pitting corrosion, crevice corrosion and trench corrosion is calculated. The occurrence frequency is divided by the total number of decision trees to obtain the corresponding prediction probability.
[0041] In an optional embodiment, when the normalized value corresponding to the corrosion volume loss in the normalized corrosion morphology topological feature vector is less than 0.01 and the normalized value corresponding to the average depth is less than 0.05, the splitting path of most decision trees in the main classifier will be directed to the pitting corrosion category, making the prediction probability of pitting corrosion higher than the other two categories.
[0042] S204: Select the label of the localized corrosion type with the highest predicted probability as the main category classification result. The main category classification result is one of pitting corrosion, crevice corrosion, or trench corrosion. In an optional embodiment, the predicted probabilities of the labels for the three types of localized corrosion—pitting corrosion, crevice corrosion, and trench corrosion—are compared, and the label for the localized corrosion type with the highest predicted probability value is determined as the primary category classification result.
[0043] In an optional embodiment, the main category classification result is used for the selection of subsequent sub-classifiers. When the main category classification result is crevice corrosion and the matrix geometry type is identified as cylindrical surface, a sub-classifier that corresponds to the combination of cylindrical surface and crevice corrosion is selected from a preset set of sub-classifiers.
[0044] S300: Based on the combination of the main category classification result and the matrix geometry type identifier, determine the target sub-classifier from the preset sub-classifier set, wherein the sub-classifier set includes a first sub-classifier corresponding to the combination of spherical surface and pitting corrosion, a second sub-classifier corresponding to the combination of cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to the combination of planar surface and trench corrosion. S301: Establish a set of sub-classifiers, which includes a first sub-classifier, a second sub-classifier, and a third sub-classifier. The first sub-classifier corresponds to the combination of spherical surface and pitting corrosion, the second sub-classifier corresponds to the combination of cylindrical surface and crevice corrosion, and the third sub-classifier corresponds to the combination of planar surface and trench corrosion. In an optional embodiment, the first sub-classifier is established as follows: residual surface data of multiple tank weld areas are collected, the matrix geometry of the tank weld areas is spherical, and the local corrosion type is confirmed by manual verification as pitting corrosion; the residual surface data are converted into corrosion morphology topological feature vectors, and the first sub-classifier is trained by using the corrosion morphology topological feature vectors as input and pitting corrosion as output label.
[0045] In an optional embodiment, the second sub-classifier is established as follows: residual surface data of the metal-cement interface region of multiple bridge pillars are collected. The surface of the bridge pillars is cylindrical, and the corrosion extends along the crevice direction. The local corrosion type is confirmed by manual verification as crevice corrosion. The residual surface data is converted into corrosion morphology topological feature vectors and trained to obtain the second sub-classifier.
[0046] In an optional embodiment, the third sub-classifier is established as follows: residual surface data of multiple flat plate connectors are collected. The substrate geometry of the flat plate connectors is planar, and the corrosion is distributed in a continuous linear pattern. The local corrosion type is confirmed by manual verification as groove corrosion. The residual surface data is converted into corrosion morphology topological feature vectors, and the third sub-classifier is trained.
[0047] S302: Combine the main category classification result with the base geometry type identifier to form a combined identifier; In an optional embodiment, when the main category classification result is pitting and the matrix geometry type is identified as spherical, "spherical-pitting" is used as the combined identifier.
[0048] In an optional embodiment, when the main category classification result is crevice corrosion and the substrate geometry type is identified as cylindrical, "cylindrical surface-crevice corrosion" is used as the combined identifier.
[0049] In an optional embodiment, when the main category classification result is trench corrosion and the substrate geometry type is identified as planar, "planar-trench corrosion" is used as the combined identifier.
[0050] S303: Match the combination identifier with the combination corresponding to each subclassifier in the subclassifier set; In an optional embodiment, the combination identifier is compared one by one with the "spherical-pitting" combination corresponding to the first sub-classifier, the "cylindrical-crevice corrosion" combination corresponding to the second sub-classifier, and the "planar-groove corrosion" combination corresponding to the third sub-classifier to determine whether there is a completely identical combination.
[0051] S304: When the combination identifier matches the combination corresponding to a certain sub-classifier, the sub-classifier is determined as the target sub-classifier; In an optional embodiment, when the combination is identified as "cylindrical surface-crevice corrosion" and is consistent with the combination corresponding to the second sub-classifier, the second sub-classifier is determined as the target sub-classifier.
[0052] In an optional embodiment, when the combination is identified as "spherical-pitted" and is consistent with the combination corresponding to the first sub-classifier, the first sub-classifier is determined as the target sub-classifier.
[0053] In an optional embodiment, when the combination is identified as “planar-trench corrosion” and matches the combination corresponding to the third sub-classifier, the third sub-classifier is determined as the target sub-classifier.
[0054] S400: Input the corrosion morphology topological feature vector into the target sub-classifier to obtain the sub-class classification result. The sub-class classification result represents the corrosion development stage under the same local corrosion type. S401: Construct a target sub-classifier. The target sub-classifier is a supervised learning model. The samples used to train the target sub-classifier include residual surface data, corresponding local corrosion type labels, and corrosion development stage labels. In an optional embodiment, the corrosion development stage label includes early stage, middle stage and late stage, wherein the early stage refers to the stage where the volume of corrosion products is less than a preset first volume threshold, the middle stage refers to the stage where the volume of corrosion products is greater than or equal to the first volume threshold and less than a preset second volume threshold, and the late stage refers to the stage where the volume of corrosion products is greater than or equal to the second volume threshold; the first volume threshold and the second volume threshold are determined based on the statistical results of the corrosion product volume at different time points in the accelerated corrosion experiment.
[0055] In an optional embodiment, the samples used to train the target sub-classifier are obtained as follows: multi-stage accelerated corrosion experiments are conducted on the same type of metal-cement composite samples, and residual surface data are obtained at corrosion times of 7 days, 14 days, and 28 days. The samples corresponding to 7 days are labeled as the initial stage, 14 days as the middle stage, and 28 days as the late stage. Combined with the confirmed local corrosion type labels (such as crevice corrosion), a complete training sample containing residual surface data, local corrosion type labels, and corrosion development stage labels is formed.
[0056] In an optional embodiment, each residual surface data is converted into a corrosion morphology topological feature vector, which includes the average depth, the arithmetic mean deviation of the profile Ra, the slope of the corrosion pit edge, and the corrosion volume loss. The corrosion morphology topological feature vector is used as the input feature, and the corresponding corrosion development stage label is used as the output label to train the target sub-classifier.
[0057] S402: Input the corrosion morphology topological feature vector into the target sub-classifier and calculate the predicted probability of each candidate corrosion development stage; In an optional embodiment, the target sub-classifier is a random forest model containing multiple decision trees. Each decision tree makes node judgments based on the feature values in the erosion morphology topological feature vector and finally outputs a erosion development stage label. The erosion development stage labels output by all decision trees are counted, and the occurrence frequency of the initial, middle and late stages is calculated. The occurrence frequency is divided by the total number of decision trees to obtain the corresponding prediction probability.
[0058] In an optional embodiment, when the corrosion volume loss in the corrosion morphology topological feature vector is large and the average depth distribution is wide, the decision paths of most decision trees in the target subclassifier will be directed to late-stage labels, making the prediction probability of late-stage labels higher than that of early-stage and mid-stage labels.
[0059] S403: Select the corrosion development stage with the highest predicted probability as the sub-category classification result. The sub-category classification result represents the corrosion development stage under the same local corrosion type. In an optional embodiment, the predicted probabilities of the three corrosion development stages—early, middle, and late—are compared, and the corrosion development stage with the highest predicted probability value is determined as the sub-category classification result.
[0060] In an optional embodiment, when the main category classification result is crevice corrosion and the target sub-classifier is the second sub-classifier, the sub-category classification result is used to characterize the development process of crevice corrosion on the cylindrical substrate, providing a basis for assessing the remaining life of the structural component.
[0061] S500: Based on the main category classification results, sub-category classification results, and corrosion volume loss in the corrosion morphology topological feature vector, generate a local corrosion assessment report for structural components.
[0062] S501: Determine the type of localized corrosion based on the main category classification results; In an optional embodiment, when the main category classification result is pitting corrosion, the local corrosion type is determined to be pitting corrosion; when the main category classification result is crevice corrosion, the local corrosion type is determined to be crevice corrosion; when the main category classification result is trench corrosion, the local corrosion type is determined to be trench corrosion.
[0063] In an optional embodiment, localized corrosion type is used to describe the distribution pattern of corrosion on the surface of the structural component, wherein pitting corrosion corresponds to isolated corrosion pits, crevice corrosion corresponds to corrosion areas extending along structural crevice, and trench corrosion corresponds to linear corrosion areas continuously distributed in a certain direction.
[0064] S502: Determine the corrosion development stage corresponding to the localized corrosion type based on the sub-category classification results; In an optional embodiment, when the subcategory classification result is initial, the corrosion development stage is determined to be initial; when the subcategory classification result is intermediate, the corrosion development stage is determined to be intermediate; and when the subcategory classification result is late, the corrosion development stage is determined to be late.
[0065] In an optional embodiment, the corrosion development stages are used to characterize the process of corrosion evolution over time, wherein the initial stage represents a stage where the corrosion products are small in volume and the corrosion pits are shallow, the intermediate stage represents a stage where the corrosion products are of medium volume and the corrosion pits are significantly expanded in depth and area, and the late stage represents a stage where the corrosion products are large in volume, the corrosion pits are connected, or the matrix is significantly thinned.
[0066] S503: Extract corrosion volume loss from corrosion morphology topological feature vector; In an optional embodiment, the corrosion volume loss is calculated by the voxel integration method, specifically: the residual surface data is divided into unit voxels in three-dimensional space, the number of all voxels with a depth less than the ideal matrix surface is counted, and multiplied by the volume of a single voxel to obtain the corrosion volume loss.
[0067] In an optional embodiment, corrosion volume loss is used to quantify the degree of material loss of the structural component in the target area and is a core parameter for evaluating the macroscopic rate of localized corrosion.
[0068] S504: Combines localized corrosion type, corrosion development stage and corrosion volume loss to generate a localized corrosion assessment report for structural components; In an optional embodiment, the localized corrosion assessment report for the structural component is output in text form, including: "Localized corrosion type: crevice corrosion; corrosion development stage: intermediate; corrosion volume loss: [specific value] cubic millimeters", where the specific value is automatically filled in by the system.
[0069] In an optional embodiment, the structural component local corrosion assessment report is used to support the durability evaluation of the metal-cement composite structure in a humid and hot environment. For example, in Embodiment 1, the corrosion behavior of the cylindrical steel foot of the ceramic insulator near the cement bonding interface is assessed in stages to provide a basis for maintenance decisions.
[0070] In an optional embodiment, the assessment report of localized corrosion of structural components can be exported as a standard data file for subsequent regression analysis to obtain the influence weights of different environmental factors on the corrosion rate.
[0071] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0072] Based on the same inventive concept, this application also provides a system for analyzing localized corrosion of complex structural components. The solution provided by this system is similar to the solution described in the above-described method. Therefore, the specific limitations of one or more embodiments of the system for analyzing localized corrosion of complex structural components provided below can be found in the limitations of the method for analyzing localized corrosion of complex structural components described above, and will not be repeated here.
[0073] In one exemplary embodiment, a system for analyzing localized corrosion of complex structural components is provided, comprising: Each module in the aforementioned system for analyzing localized corrosion of complex structural components can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.
[0074] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 2As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for analyzing localized corrosion of complex structural components. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.
[0075] Those skilled in the art will understand that Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0076] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.
[0077] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.
[0078] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.
[0079] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0080] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0081] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0082] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method of analyzing local corrosion of a complex structure, characterized by, The method comprises the following steps: acquiring original three-dimensional surface data of a target region of a complex structural component exposed to a corrosive environment, and extracting a substrate geometric type identifier and a corrosion topography topological feature vector from the original three-dimensional surface data; inputting the corrosion topography topological feature vector into a main classifier to obtain a main category classification result, the main category classification result being a local corrosion type label, and the local corrosion type label including pitting corrosion, crevice corrosion or groove corrosion; determining a target sub-classifier from a preset sub-classifier set according to a combination of the main category classification result and the substrate geometric type identifier, wherein the sub-classifier set includes a first sub-classifier corresponding to a spherical surface and pitting corrosion, a second sub-classifier corresponding to a cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to a plane and groove corrosion; inputting the corrosion topography topological feature vector into the target sub-classifier to obtain a sub-category classification result, the sub-category classification result representing a corrosion development stage under the same local corrosion type; generating a structural component local corrosion evaluation report based on the main category classification result, the sub-category classification result, and a corrosion volume loss in the corrosion topography topological feature vector.
2. The method of claim 1, wherein: The method further comprises the following steps for extracting the substrate geometric type identifier and the corrosion topography topological feature vector: optically scanning a target region of a complex structural component exposed to a corrosive environment to obtain original three-dimensional surface data; determining a substrate geometric type corresponding to the target region according to structural features of the complex structural component, and generating a substrate geometric type identifier, the substrate geometric type identifier being one of a spherical surface, a cylindrical surface or a plane; preprocessing the original three-dimensional surface data to remove scanning noise and invalid point clouds, and obtaining valid three-dimensional surface data; extracting a corrosion topography topological feature vector from the valid three-dimensional surface data, the corrosion topography topological feature vector including an average depth, an arithmetic mean deviation Ra of a profile, a corrosion pit edge slope and a corrosion volume loss.
3. The method of claim 2, wherein: The method further comprises the following steps for obtaining the main category classification result: constructing a main classifier, the main classifier being a supervised learning model, and training the main classifier using samples including residual surface data and corresponding local corrosion type labels; performing normalization processing on the corrosion topography topological feature vector to unify numerical ranges of each dimension of the corrosion topography topological feature vector to the interval of 0 to 1; inputting the normalized corrosion topography topological feature vector into the main classifier to calculate prediction probabilities of three types of local corrosion type labels, i.e., pitting corrosion, crevice corrosion and groove corrosion; selecting a local corrosion type label with the maximum prediction probability as the main category classification result.
4. The method of claim 3, wherein: The method further comprises the following steps for determining the target sub-classifier from the preset sub-classifier set: establishing a sub-classifier set, the sub-classifier set including a first sub-classifier, a second sub-classifier and a third sub-classifier, the first sub-classifier corresponding to a combination of a spherical surface and pitting corrosion, the second sub-classifier corresponding to a combination of a cylindrical surface and crevice corrosion, and the third sub-classifier corresponding to a combination of a plane and groove corrosion; combining the main category classification result and the substrate geometric type identifier to form a combined identifier; matching the combination identifier with a combination corresponding to each sub-classifier in the sub-classifier set; when the combination identifier is consistent with the combination corresponding to a sub-classifier, determining the sub-classifier as a target sub-classifier.
5. The method of claim 4, wherein: the obtained sub-category classification result comprises: constructing a target sub-classifier, the target sub-classifier being a supervised learning model, and the sample used for training the target sub-classifier comprising residual surface data, corresponding local corrosion type label and corrosion development stage label; inputting the corrosion topography topological feature vector into the target sub-classifier to calculate the prediction probability of each candidate corrosion development stage; selecting the corrosion development stage with the maximum prediction probability as the sub-category classification result, the sub-category classification result representing the corrosion development stage under the same local corrosion type.
6. The method of claim 5, wherein: the generated structure local corrosion evaluation report comprises: determining the local corrosion type according to the main category classification result; determining the corrosion development stage corresponding to the local corrosion type according to the sub-category classification result; extracting the corrosion volume loss from the corrosion topography topological feature vector; combining the local corrosion type, the corrosion development stage and the corrosion volume loss to generate a structure local corrosion evaluation report.
7. The method of claim 6, wherein: before the corrosion topography topological feature vector is extracted, further comprising: according to the base body geometric type identifier, selecting a corresponding mathematical model of curved surface to fit the original three-dimensional surface data to generate an ideal base body curved surface, wherein when the base body geometric type identifier is a spherical surface, a least square method is used to fit a spherical surface equation, and when the base body geometric type identifier is a cylindrical surface, a least square method is used to fit a cylindrical surface equation; point-by-point difference value is performed between the original three-dimensional surface data and the ideal base body curved surface to obtain residual surface data; depth threshold segmentation is performed on the residual surface data, and a region with a depth greater than a corrosion threshold is extracted as an effective corrosion region; the corrosion topography topological feature vector is extracted based on the residual surface data of the effective corrosion region.
8. A complex structure local corrosion analysis system using the complex structure local corrosion analysis method according to any one of claims 1 to 7, characterized by comprising: a data acquisition module configured to acquire original three-dimensional surface data of a target region of a complex structure subjected to an erosion environment, and extract a base body geometric type identifier and a corrosion topography topological feature vector from the original three-dimensional surface data; a main classification module configured to input the corrosion topography topological feature vector into a main classifier to obtain a main category classification result, the main category classification result being a local corrosion type label, the local corrosion type label comprising pitting, crevice corrosion or groove corrosion; a sub-classification selection module configured to determine a target sub-classifier from a preset sub-classifier set according to a combination of the main category classification result and the base body geometric type identifier, wherein the sub-classifier set comprises a first sub-classifier corresponding to a spherical surface and pitting, a second sub-classifier corresponding to a cylindrical surface and crevice corrosion, and a third sub-classifier corresponding to a plane and groove corrosion; a development stage discrimination module configured to input the corrosion topography topological feature vector into the target sub-classifier to obtain a sub-category classification result, the sub-category classification result representing a corrosion development stage under the same local corrosion type; An evaluation report generation module is configured to generate a local corrosion evaluation report of the structure based on the main category classification result, the sub-category classification result, and the corrosion volume loss in the corrosion topography feature vector. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The computer program is executed by the processor to implement the steps of the complex structure local corrosion analysis method in any one of claims 1 to 7.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the complex structure local corrosion analysis method in any one of claims 1 to 7.