A method and system for detecting quality defects in a building material

Abstract

The application discloses a kind of building material quality defect detection method and system, the method comprises: obtaining the steel member material image under the tension stress of each time point collection, the steel member material image of each steel member material image is preprocessed, the steel member body data of the material image to be measured of each time point and steel member is obtained;The crack region of the material image to be measured of each time point is identified, and the edge point of crack region is screened starting edge point and end edge point;According to starting edge point and end edge point, the turning point of crack is screened, the crack sawtooth degree between each time point is calculated by the turning point of crack, the crack propagation change amount is calculated by the crack sawtooth degree between each time point crack;Through crack propagation change amount, judge the fatigue detection result of the steel member material image under the tension stress.

Description

Technical Field

[0001] This invention relates to the field of building material quality, and in particular to a method and system for detecting quality defects in building materials. Background Technology

[0002] As building structures continuously bear various loads and environmental effects during long-term use, steel structural components, in particular, gradually exhibit material fatigue. The initiation and propagation of fatigue cracks are one of the key reasons for the performance degradation and even failure of building components. Therefore, accurately detecting the fatigue state of components is of great significance for extending the service life of buildings and ensuring structural safety. Traditional fatigue detection methods mostly rely on physical and mechanical tests in laboratories or use technologies such as ultrasound and X-rays to detect the presence and development of cracks. These methods typically suffer from problems such as long detection cycles, high costs, and stringent requirements for the professional skills of technicians. Against this backdrop, developing an image-based fatigue characteristic detection method is particularly important.

[0003] In modern construction engineering, tensile stress testing is widely used to evaluate the fatigue performance of building materials. In the laboratory, by acquiring surface image data of building components under different levels of tensile stress, it is possible to effectively capture the deformation and damage evolution of steel structure building components under stress, especially the generation and propagation behavior of microcracks.

[0004] Further investigation revealed that traditional image processing methods, such as those using Convolutional Neural Networks (CNNs), which automatically learn features directly from the original image and extract high-level features through multi-layer convolution and pooling operations to output detection results, still have certain limitations when processing complex crack morphologies. In practical applications, cracks not only exist in a single location but may also form dense crack clusters across the entire surface or in local areas. Crack density has a significant impact on the fatigue performance of materials; highly dense crack clusters often indicate that the material has entered a critical state of fatigue failure. Therefore, traditional image detection methods struggle to effectively distinguish between crack density and its propagation behavior (and thus cannot provide valuable assessments based on the complex morphology of cracks), resulting in inaccurate crack density assessments. Summary of the Invention

[0005] This invention provides a method for detecting quality defects in building materials, comprising the following steps:

[0006] Acquire images of steel components under tensile stress at each acquisition moment, preprocess each image to obtain the image of the material to be tested at each moment and the body data of the steel component from the image of the steel component material;

[0007] The crack region is identified in the image of the material under test at each time moment, and the starting edge point and ending edge point of the crack region are screened. The turning point of the crack is screened according to the starting edge point and ending edge point. The crack serration degree between cracks at each time moment is calculated through the crack turning point. The crack propagation change is calculated through the crack serration degree between cracks at each time moment.

[0008] The fatigue test results of the steel component material image under tensile stress are judged by the crack propagation change.

[0009] A preset crack propagation change threshold y is used to determine whether the crack propagation change is greater than or equal to the crack propagation change threshold y; if so, it is determined that the steel component in the material image of the steel component has fatigue risk.

[0010] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0011] Analysis of the above-mentioned method and system for detecting quality defects in building materials provided by the present invention shows that, in specific applications, edge detection algorithms (such as Sobel, Canny, etc.) are used to identify crack regions in the material image at each moment, extract the edge points of the crack contour, and form an edge point sequence. The above method reduces data complexity by identifying the edge regions of the crack. Then, the Douglas-Puk algorithm is used to determine the contour of the crack region using a sliding window to filter the starting and ending edge points. Simultaneously, the area of ​​the connected integral is calculated using the sliding window, and local maxima within the area are selected as positive value points of the connected surface. These two points are then connected by a straight line to obtain a point-line. Next, the perpendicular distances from all intermediate edge points to the point-line are calculated, and the intermediate edge points deviating from the main path—that is, the edge points with the largest perpendicular distances—are selected as extreme points of perpendicular distance. These extreme points reflect the most convex or concave positions that the crack may correspond to. The coincidence of the positive value points of the connected surface and the extreme points of perpendicular distance is verified to identify turning points, helping to identify significant changes or potential turning points in the crack path. In other words, by calculating the area of ​​the connected integral, local maxima are selected as positive value points of the connected surface, and further, extreme points of perpendicular distance are determined through perpendicular distance calculation. The coincidence of these two points is verified and used as the crack turning point.

[0012] Crack serration is calculated based on the coordinates, distance, and direction angle of the inflection point to measure crack irregularity and classify crack types accordingly. By calculating the serration density of various crack types, high-fatigue-risk cracks are identified, and a crack database is constructed. Simultaneously, for the crack set with the minimum serration density, a skeleton algorithm is used to extract the crack tip, and a stress intensity factor is calculated using a stress gradient algorithm. Finally, the stress intensity factor is used to assess crack propagation trends and failure risks to determine the crack characteristics of building components. Attached Figure Description

[0013] Figure 1 This is a flowchart of a method for detecting quality defects in building materials, as described in Example 1.

[0014] Figure 2 This is a flowchart illustrating the crack propagation variation of a building material quality defect detection method according to Example 1.

[0015] Figure 3 This is a schematic diagram illustrating the inflection point screening process in a method for detecting quality defects in building materials, as described in Example 1.

[0016] Figure 4 This is a flowchart illustrating the calculation of crack propagation variation in a method for detecting quality defects in building materials, as described in Example 1.

[0017] Figure 5 This is a schematic diagram of the inflection point direction angle of a method for detecting quality defects in building materials according to Embodiment 1;

[0018] Figure 6 This is a flowchart illustrating the crack propagation variation calculated using the stress intensity factor in a method for detecting quality defects in building materials, as described in Example 1.

[0019] Figure 7 This is a flowchart illustrating the acquisition of the high-strength stress factor in a method for detecting quality defects in building materials, as described in Example 1.

[0020] Figure 8 This is a schematic diagram showing the coordinates of the crack tip point in a method for detecting quality defects in building materials according to Example 1.

[0021] Figure 9 This is a flowchart illustrating the calculation of the stress factor in a method for detecting quality defects in building materials, as described in Example 1.

[0022] Figure 10 This is a flowchart of a building material quality defect detection system according to Example 2;

[0023] Labels: Acquisition module 10; Analysis module 20; Judgment module 30. Detailed Implementation

[0024] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0026] Example 1

[0027] like Figure 1 As shown, this application provides a method for detecting quality defects in building materials, including the following steps:

[0028] S1: Acquire images of steel components under tensile stress at each time point (i.e., multiple time point detection and acquisition images), preprocess each steel component material image to obtain the image of the material to be tested at each time point and the steel component body data of the steel component material image;

[0029] It should be noted that collecting image data of steel components under different tensile stresses is to capture the changes in the state of the steel components under load; image preprocessing of steel components is to ensure image quality, eliminate noise, and improve image resolution and contrast. Common preprocessing methods include noise reduction, smoothing, edge detection, etc., which help improve the accuracy of subsequent crack identification.

[0030] Meanwhile, the preprocessing also yielded the steel component material data, including the material composition, shape, thickness, and other data of the steel component itself. These data all affect the cracking under tensile stress and can reflect the actual fatigue condition of the steel component material, especially the small changes in the early stage of cracking.

[0031] S2: Identify crack regions in the image of the material under test at each time step, and filter the starting and ending edge points of the crack regions; filter the inflection points of the cracks based on the starting and ending edge points, calculate the crack serration degree between cracks at each time step through the crack inflection points, and calculate the crack propagation change amount based on the crack serration degree between cracks at each time step.

[0032] It should be noted that cracks are an important indicator of material fatigue. Identifying crack regions through image processing of the material under test can help us find potential fatigue cracks in the material.

[0033] The turning point of a crack is usually a key point where the crack morphology undergoes a significant change in trajectory (i.e., a key pixel, marking the direction and location of the crack). A parallelogram is formed by connecting the starting and ending edge points of the crack region's outline with line segments, and then perpendicularly connecting the intermediate edge points to the line segments between the starting and ending edge points. The intermediate edge points are the edge pixels located in the edge point sequence between the starting and ending edge points. The turning point is identified by the area of ​​the parallelogram formed by the intermediate edge points. The area of ​​the parallelogram represents the relationship between crack displacement and force (i.e.,...). This refers to crack growth. Alternatively, in a two-dimensional coordinate system, the area of ​​a parallelogram can represent the amount of change of a variable over time. When the area of ​​the parallelogram is particularly large, it may be due to changes in external forces or physical states, indicating that the system is experiencing a turning point. The crack serration is calculated through the turning point, thus providing an important indicator for describing the irregularity of the crack edge. The crack serration is usually related to the complexity of local stress concentration; the greater the serration, the higher the risk of crack propagation. Therefore, collecting the aforementioned time-varying image feature data and then calculating the crack serration can help predict the likelihood of crack propagation.

[0034] By calculating the change in crack serration at different time points, the change in crack propagation can be estimated, thereby reflecting whether the fatigue of the steel component is significant.

[0035] S3: The fatigue detection result of the steel component material image under tensile stress is judged by the crack propagation change;

[0036] A preset crack propagation change threshold y is set, and it is determined whether the crack propagation change is greater than or equal to the crack propagation change threshold y; if so, it is determined that the steel component in the material image of the steel component has fatigue risk.

[0037] It should be noted that the calculation of crack propagation change provides a direct basis for judging the fatigue life of steel components. The faster the crack propagation change, the higher the fatigue risk faced by the steel component. Therefore, evaluating fatigue test results by crack propagation change can accurately reflect the fatigue performance of materials under specific loads. By evaluating crack propagation change, the fatigue life of materials under long-term stress can be predicted. If the crack propagation change is large, the service life of the steel component is short and it may fail in a short time. Conversely, a slow crack propagation change means that the fatigue life of the steel component is long, ensuring timely maintenance or replacement to avoid failure.

[0038] Specifically, such as Figure 2As shown, in step S2, crack regions are identified in the image of the material under test at each time step, and the inflection points of the crack regions are screened. The crack serration degree between cracks at each time step is calculated based on the crack inflection points, and the crack propagation change is calculated based on the crack serration degree between cracks at each time step. The specific operation steps are as follows:

[0039] S21: Use edge detection algorithms (such as Sobel, Canny, etc.) to identify crack regions in the material image at each time step; extract edge points from the contour of the crack region to form an edge point sequence at multiple time steps;

[0040] It should be noted that the edge detection algorithm is used to identify the contour region of the crack in the image of the material under test, and the edge points are found through the contour region to obtain the edge point sequence, thereby reducing data complexity. The edge points are used to focus on the key points for extracting the crack shape in subsequent steps.

[0041] At each time point, edge detection algorithms (such as Sobel, Canny, etc.) are applied to the image of the material under test to identify the edges of the cracks. The edge detection algorithm can find areas in the image of the material under test where the pixel values ​​change significantly, ensuring that the selected crack area is accurate and providing reliable crack contour data for subsequent processing. This usually corresponds to the outline of the crack. The extraction of crack areas is common knowledge and will not be elaborated here.

[0042] S22: Use the Douglas-Puk algorithm to randomly select an edge point from the sequence of edge points of the crack region at each time step, and use it as the starting edge point; use the starting edge point to form a sliding window for the contour of the crack region;

[0043] For each sliding window, count the number of edge points, find the edge point within the sliding window that covers 1 edge point, and calculate the straight-line distance between the starting edge point and the edge point within the sliding window that covers 1 edge point; select the maximum straight-line distance as the ending edge point of the crack region's contour.

[0044] It should be noted that the Douglas-Puk algorithm is used to simplify the curve and determine the key boundary points (i.e., the start and end edge points) of the crack region. Because the outline of the crack region is circular or rectangular, regardless of its shape or regularity, the outline is always closed. Therefore, a random edge point is selected as the start edge point, and the sliding window slides along the start edge point across all edge points of the crack region's outline, allowing for the counting of edges within the sliding window. Furthermore, since the end of the crack is the endpoint of its propagation, the end edge point will only have one neighboring pixel, meaning there can only be one pixel within the sliding window. Also, because the start and end points of the crack are far apart, the edge point with the furthest distance (number of edges) of 1 is selected as the end edge point of any possible crack region's outline.

[0045] S23: Connect the starting edge point and the ending edge point with a straight line to obtain a point-line;

[0046] For each sliding window, select sliding windows that are parallel to the line of the point, and calculate the sum of non-zero pixels for the parallel sliding windows to obtain the area of ​​the connected integral.

[0047] Find the local maximum points of the edge points in the connected integral area and use them as the maximum points of the connected surface.

[0048] It should be noted that the Douglas-Pock algorithm is a commonly used fitting algorithm used to find the straight line segment representing the curve among a set of edge points. The Douglas-Pock algorithm is used to identify the start and end points of the crack region and connect these points with straight lines. By identifying the main edges of the crack, it can help subsequent steps to accurately locate the changes in the crack, especially by identifying the turning areas of the crack by determining the straight line segments.

[0049] The sliding window is parallel to the line of points, that is, the direction of the sliding window is consistent with the main direction of the cracks that may exist on the line of points, so as to avoid false judgments caused by window tilt; each window is binarized, and the area of ​​the connected integral is calculated by counting the non-zero pixels. The area is used to represent the geometric irregularity of the cracks that may be captured.

[0050] At the same time, by filtering each edge point in the connected integral area and comparing the gray values ​​of adjacent edge points, edge points with higher values ​​in the surrounding neighborhood are selected as local maximum points and used as positive value points of the connected surface. This can help to understand the most unstable points in the area, that is, the deviating edge points.

[0051] S24: Filter the middle edge points of the outline of the crack region, and calculate the perpendicular distance of the straight line to each middle edge point.

[0052] The edge points with the largest vertical distances are selected as the vertical distance extreme points.

[0053] It should be noted that, starting from the straight line between the start and end edge points of the crack (obtained using the Douglas-Puk algorithm), the vertical distance from each intermediate edge point in the crack profile to that straight line is calculated. Calculating the vertical distance helps identify points that significantly deviate from the main crack path. Through the vertical distance, the morphological changes of the crack can be quantified, and the trend of crack change can be further identified (i.e., the calculation of the vertical distance can reveal the degree to which intermediate edge points deviate from the straight line (i.e., the baseline); the extreme points of the vertical distance are the peaks in the vertical distance of all edge points that are significantly higher than the surrounding values, which are the most convex or concave positions that the crack may correspond to).

[0054] S25: Verify the coincidence of the extreme point of the connected surface with the extreme point of the vertical distance. When the two coincide, it is taken as the turning point.

[0055] It should be noted that both the positive value point of the connected surface and the extreme value point of the vertical distance are considered as points of potential deviation, i.e., the sawtooth points of the crack. When the coordinates of the positive value point of the connected surface passing through both coincide with the extreme value point of the vertical distance, this edge point is determined as a turning point, such as... Figure 3 As shown;

[0056] S26: Calculate the crack serration degree between adjacent inflection points, classify cracks by the crack serration degree between adjacent inflection points, and obtain different types of crack sets at each time point; calculate the serration density in each type of crack set by the number of cracks and the number of inflection points in the crack set; calculate the crack propagation change by the serration density at each time point.

[0057] It should be noted that the serration density obtained by calculating the number of cracks and the number of inflection points in different types of crack sets can provide information about crack distribution density and reveal the concentration trend of crack development.

[0058] Based on the above analysis, calculating the serration density at each moment to assess the change in crack propagation helps to understand the crack propagation rate over time and provides an accurate understanding of the material failure time.

[0059] Specifically, such as Figure 4 As shown, in step S26, the crack serration degree between adjacent inflection points is calculated. Cracks are classified based on this serration degree to obtain different types of crack sets at each time point. The serration density of each type of crack set is calculated using the number of cracks and inflection points in the crack sets. The crack propagation change is calculated using the serration density at each time point. The specific operation steps are as follows:

[0060] S261: Determine the coordinates of the turning point at each moment, calculate the direction angle of adjacent turning points using the coordinates of each turning point, and calculate the crack serration degree between adjacent turning points at each moment based on the direction angle of the adjacent turning points. The crack serration degree is:

[0061] ;

[0062] in, This is represented by the crack serration degree at the i-th inflection point;

[0063] The coordinates of the i-th turning point;

[0064] The coordinates of the v-th turning point;

[0065] Let be the distance between the i-th turning point and the v-th turning point;

[0066] Let be the directional angle between the i-th turning point and the v-th turning point;

[0067] N represents the number of inflection points;

[0068] To take the absolute value;

[0069] It should be noted that, firstly, the “inflection point” at each moment, that is, the edge point of the crack, is determined, and then the directional angle between adjacent inflection points is calculated. The changing trend of the crack profile is measured by these directional angles. The crack serration degree (or crack complexity) is defined as a certain “irregularity” or “serrated shape” between the coordinates of each inflection point.

[0070] Crack serration reflects the complexity of crack morphology. A higher crack serration usually means that the crack edge is more irregular, which may be due to local stress concentration or material heterogeneity. Crack serration is quantified by the coordinates of the inflection points and the distances and directional angles between them. In this way, the changing angles and distances between edge points are calculated, thus measuring the variability of the crack edge. The higher the crack serration, the more irregular the crack edge changes and the more severe the local stress concentration. Generally, a larger serration may indicate a higher degree of stress concentration in the crack region, which may lead to faster crack propagation. Therefore, crack serration is crucial for predicting crack growth rate and fatigue life.

[0071] S262: Based on the crack serration degree, set the crack serration degree classification threshold range q, determine the range of the crack serration degree at each moment within the crack serration degree classification threshold range q, obtain multiple types of cracks at each moment, and set each type of crack at each moment to obtain a set of multiple types of cracks.

[0072] It should be noted that cracks with higher serrations typically exhibit irregular shapes, and these cracks may be more prone to propagation and pose a higher fatigue risk; while cracks with lower serrations may be more regular, propagate more slowly, and have a lower fatigue risk. Classifying crack serrations and setting a threshold range helps distinguish different types of cracks. This means classifying cracks into different categories based on their serration size. For example, some cracks have lower serrations, resulting in more regular crack shapes and slower fatigue propagation; others are more complex, with higher serrations, irregular crack shapes, and a higher fatigue risk. Each crack type is classified based on its serration and threshold. Crack type classification helps in... Analyzing the varying complexities of cracks allows for the assignment of different fatigue risk levels to different types of cracks. Furthermore, by grouping cracks of each classified type into sets (i.e., crack sets), which group similar cracks based on their initiation points or crack tips, the location and distribution of different crack types on the surface of the material or steel component can be more clearly understood. For example: Crack set 1: a set of all low-serration cracks with relatively regular crack paths; Crack set 2: a set of all high-serration cracks with more complex crack morphologies. By connecting the crack tips of these sets, the crack inflection points and fracture paths can be identified, thus forming a complete crack pattern (i.e., a crack set).

[0073] By classifying cracks, high-risk cracks (such as cracks with large serrations) can be identified. These cracks are more likely to propagate and therefore have a greater impact on the fatigue life of steel components. By classifying crack types, the development trend of cracks can be predicted more accurately, and targeted preventive measures can be taken.

[0074] S263: For each type of crack set at each time step, calculate the number of cracks and the number of inflection points in that type of crack set. Then, use the number of cracks and the number of inflection points to calculate the serration density in each type of crack set at each time step. The serration density is:

[0075]

[0076] K represents the serration density, reflecting the complexity and degree of variation of the crack profile;

[0077] e is denoted as the natural constant;

[0078] P represents the entropy of the crack region's contour, reflecting the complexity of the information in the image;

[0079] Dmax represents the maximum directional angle of the inflection point of the crack set in the crack region profile (e.g., the maximum vertical distance from the inflection point to its fitted baseline, i.e., the inflection point is finally selected by the vertical distance of the calculated line of the edge point in step S2, and the inflection point is also on the vertical distance).

[0080] Dmin represents the minimum directional angle of the inflection point of the crack set in the crack region profile (e.g., the minimum perpendicular distance from the inflection point to its fitted baseline, such as...). Figure 5 (as shown)

[0081] This represents the number of cracks in the i-th crack set;

[0082] This is expressed as the number of inflection points;

[0083] Dj represents the distance from the j-th inflection point to the straight line (such as the distance between the inflection point and the start and end edge points), used to quantify the direction angle of the inflection point;

[0084] To explain, traditional methods extract image features into feature vectors, while convolutional neural networks (CNNs) automatically learn features directly from the original image, extracting high-level features through multiple convolution and pooling operations, without the need for manual feature design.

[0085] The parameters designed in this application embodiment, such as sawtooth density, entropy, direction angle of inflection point, number of cracks, number of inflection points, Dj distance, etc., are mostly manually designed features. That is, specific geometric or statistical features are extracted from the image, and then these features are combined into vectors and input into a classifier or regression model for recognition processing, and then sawtooth density, etc. are calculated.

[0086] By calculating the relationship between the number of cracks and the number of inflection points in each crack type, the serration density is further calculated. The calculation of serration density involves multiple factors, and the specific formula comprehensively considers the entropy of the crack profile, the maximum and minimum directional angles of the inflection points, and the distance between the inflection points.

[0087] The serration density reflects the complexity and variability of the crack profile; by calculating the serration density, the stress concentration and potential crack propagation variation in the crack region can be further quantified.

[0088] Entropy (P) represents the information complexity of the crack region. The higher the entropy value, the more complex the crack profile, which usually means that the crack propagation is more difficult. Furthermore, the maximum and minimum directional angles (Dmax, Dmin) of the inflection points reflect the degree of abrupt change in crack morphology. The larger the change angle of the inflection points, the higher the possibility and speed of crack propagation. At the same time, by calculating these values, the relationship between the number of cracks and the number of inflection points can be obtained as a quantitative index reflecting crack morphology and its evolution, namely, serration density.

[0089] Cracks with high serration density indicate a more complex crack region, which usually suggests a higher fatigue risk in that region. This complex crack morphology helps us predict the crack propagation rate and thus assess the remaining service life of steel components.

[0090] S264: Select the crack set with the largest serration density through the serration density as high fatigue risk cracks; construct a crack database for the crack set of high fatigue risk cracks at each time moment; calculate the crack propagation change using the crack set of high fatigue risk cracks at each time moment in the crack database.

[0091] It should be noted that, finally, by selecting the set of cracks with the highest serration density (i.e. high fatigue risk cracks), a crack database is constructed. The crack database contains information on high-risk cracks at each moment and is used to calculate the crack propagation change.

[0092] By focusing on high-fatigue-risk cracks and recording their propagation behavior, more accurate detection results can be provided for predicting the fatigue life of the entire steel structure; by calculating the crack propagation variation, the crack growth trend can be accurately predicted.

[0093] Research has revealed that crack propagation rate and direction have a significant impact on the performance and safety of steel components. By calculating the change in crack propagation, the crack propagation trend can be understood, thereby identifying potential fatigue points in advance. The crack propagation rate is directly related to the fatigue life of the material or steel component. As the crack grows larger during propagation, it may eventually lead to the failure of the steel component. Furthermore, the stress intensity factor (SIF) is introduced when calculating the crack propagation rate. The SIF is one of the core parameters of crack propagation; it measures the stress intensity at the crack tip and is usually one of the driving forces of crack propagation. The SIF directly reflects the local stress state at the crack tip; the more concentrated the stress at the crack tip, the greater the likelihood of crack propagation. By measuring the SIF at the crack tip, the crack propagation rate and trend can be predicted.

[0094] Specifically, such as Figure 6As shown, in step S264, the crack set with the highest serration density is selected as the high fatigue risk crack; a crack database is constructed for the high fatigue risk crack set at each time step; the crack propagation change is calculated using the high fatigue risk crack set at each time step in the crack database. The specific operation steps are as follows:

[0095] S2641: Select the set of cracks with the largest sawtooth density at each moment based on the sawtooth density, and identify them as high fatigue risk cracks.

[0096] It should be noted that the set of cracks with the highest serration density at each moment is selected as high fatigue risk cracks; by calculating the serration density, the more complex parts of the crack edge can be identified.

[0097] Cracks with high serration density usually indicate more complex crack morphology, higher local stress concentration, and are more susceptible to propagation under load. Therefore, these cracks pose a higher fatigue risk and should be prioritized for identification. This helps to focus on the most threatening areas in fatigue analysis, thereby providing a more targeted assessment.

[0098] S2642: Store the set of high fatigue risk cracks at each moment to establish a crack database;

[0099] It should be noted that the set of high fatigue risk cracks at each moment is stored in the crack database; the crack database stores the geometric features, serration density and related fatigue information of the high risk cracks at each moment, which becomes the basis for subsequent analysis and prediction of crack propagation.

[0100] Crack databases can provide data support for long-term monitoring and multiple analyses; by storing crack characteristics at different time points, the evolution of cracks can be better tracked, and a reliable basis can be provided for subsequent fatigue prediction.

[0101] S2643: Use the skeleton algorithm to filter the set of cracks with the smallest serration density type at each time step, and perform skeleton analysis on the pixels to obtain candidate tip points; obtain crack tip points based on the candidate tip points with the same coordinates of the set of cracks with the smallest serration density type at each time step; calculate the stress intensity factor at each time step using the crack tip points.

[0102] It should be noted that the skeleton algorithm is used to select the set of cracks with the minimum serration density at each time step, and candidate tip points are obtained by analyzing the pixels through the skeleton. By further filtering the crack tip points from these candidate tip points, the stress intensity factor at each time step can be calculated. This factor is a key parameter for crack propagation.

[0103] The skeleton algorithm can effectively identify crack tips by extracting the skeleton lines of cracks (i.e., the main body of cracks). Crack tips are the key locations for crack propagation, and accurate calculation of the stress intensity factor at the crack tip can provide more accurate fatigue detection results for steel components in analyzing crack propagation.

[0104] The stress intensity factor is one of the core factors affecting crack propagation. By measuring the stress intensity factor at the crack tip, the crack propagation trend and speed can be predicted. A larger stress intensity factor usually means a greater potential for crack propagation. Therefore, the influence of the stress intensity factor on cracks needs to be considered.

[0105] S2644: Calculate the crack propagation change of the crack at each moment using the set of high-fatigue-risk cracks in the crack database. The calculation formula is as follows:

[0106] ;

[0107] In the formula, Indicates at time The change in crack propagation after normalization;

[0108] This indicates that the crack database is at time [time]. In the process, the maximum serration density of the high fatigue risk crack set obtained through screening reflects the local stress concentration characteristics and complexity of the crack region at that moment;

[0109] Represents the first time interval adjacent to time t in the crack database. The maximum serration density of the set of high fatigue risk cracks at each time point;

[0110] This indicates the number of moments used for comparison in the crack database, i.e., the total number of recorded crack data points;

[0111] This represents the average serration density in the set of high fatigue risk cracks within the crack database;

[0112] Indicates at time The degree of dispersion reflected by the crack set (e.g., crack length distribution, shape statistics, etc.).

[0113] This indicates the average dispersion of the crack set within the crack database;

[0114] This represents the reference weight, used to adjust the influence of factors related to feature proportion and dispersion.

[0115] Indicates at time The stress intensity factor measured below reflects the stress condition at the crack tip and the effect of the load at that moment;

[0116] Indicates the reference or baseline stress intensity factor, used for normalization. The value can be taken as the material's threshold or the design standard value;

[0117] This represents the geometric correction factor (i.e., the steel component body data is used as a geometric correction factor, representing an exponential parameter related to material properties and shape, reflecting the sensitivity of crack propagation variation to changes in stress intensity factor).

[0118] Norm[•] represents the normalization function, used to adjust the calculated crack propagation variation to a predetermined numerical range (e.g., ...). (or other target ranges), to facilitate numerical comparison and engineering applications;

[0119] It should be noted that the crack propagation variation is calculated by using the set of high fatigue risk cracks in the crack database and other relevant information. The formula involves multiple factors, such as the maximum serration density, the average serration density, the dispersion, and the stress intensity factor. These parameters work together to determine the magnitude of the crack propagation variation and improve the accuracy of the crack propagation variation calculation.

[0120] Crack propagation variation helps predict crack propagation rate. By integrating multiple factors (such as crack geometry, local stress, stress intensity factor, etc.), the crack propagation rate can be calculated more accurately, thereby assessing the fatigue life of steel components.

[0121] The degree of dispersion indicates the range of crack morphology variation. Cracks with a larger degree of dispersion may be more unstable and have a higher propagation rate. The geometric correction factor takes into account factors such as the shape and material properties of the steel component, and helps to more accurately reflect the influence of the steel component properties on crack propagation.

[0122] Calculating the crack propagation rate can predict crack growth at different time points and help assess the remaining service life of steel components; a high propagation rate indicates that the crack has a significant impact on the fatigue of the steel component.

[0123] Studies have found that crack propagation in engineering structures is typically influenced by the stress intensity factor; the larger the stress intensity factor at the crack tip, the higher the likelihood of crack propagation. Therefore, the propagation stress intensity factor helps to more accurately predict the crack propagation process, thereby providing an assessment of structural safety. By capturing the stress field at the crack tip at each moment and calculating the stress intensity factor, the stress state changes at the crack tip can be dynamically assessed. The stress at the crack tip changes as the crack propagates, and this change can provide information on the crack propagation trend, rate, and possible failure modes. Therefore, the propagation stress intensity factor helps to understand the stress distribution of the crack at different time points in detail.

[0124] Specifically, such as Figure 7 As shown, in step S2643, the skeleton algorithm is used to filter the set of cracks with the smallest serration density type at each time step, and the skeleton analysis pixels are used to obtain candidate tip points; based on the candidate tip points with the same coordinates of the set of cracks with the smallest serration density type at each time step, the crack tip points are obtained; the stress intensity factor at each time step is calculated using the crack tip points. The specific operation steps are as follows:

[0125] S26431: Use the skeleton algorithm to filter out the crack set with the smallest sawtooth density type at each moment;

[0126] Skeleton analysis is performed on the crack set of the smallest serration density type, and the non-zero pixels in the eight neighborhood directions of each pixel in the crack set of the smallest serration density type are statistically analyzed.

[0127] By counting the number of non-zero pixels for each pixel, all pixels with a count equal to 1 are selected as candidate tip points for the crack set with the smallest serration density at each moment.

[0128] It should be noted that the set of cracks with the lowest serration density is selected, which means selecting the set of cracks with the least significant serration undulation (lowest density) from all crack types; such cracks are usually closer to ideal straight cracks, which are in line with the theoretical assumptions of linear elastic fracture mechanics (LEFM), and complex branching or highly tortuous cracks are excluded.

[0129] The selected crack set is skeletonized. The number of non-zero pixels in the 8-neighborhood of each skeleton pixel (i.e., pixel point) in the crack set with the minimum serration density is counted. The crack tip point can be determined by having only one pixel in its neighboring domain or only one endpoint at the end of the crack. Since the skeleton line of the crack set is continuous, the endpoint must connect at least one pixel point. Therefore, both the crack tip point at the beginning and the endpoint at the end of the crack set can be used as candidate tip points, avoiding reliance on complex calculations such as curvature (i.e., in step S2, the starting edge point and the ending edge point are obtained, which are not true crack tip points, but are selected from the contour of the crack region. Furthermore, the starting edge point is randomly selected, so it is not a true crack tip point).

[0130] S26432: Determine the two-dimensional coordinates of the candidate tip points of the crack set with the smallest serration density type at each time moment, and determine whether the two-dimensional coordinates of the candidate tip points of the crack set with the smallest serration density type at each time moment are the same.

[0131] If not, then the two-dimensional coordinates of the candidate tip point have a single coordinate, and the candidate tip point with all single coordinates is determined to be the end point of the crack set.

[0132] If so, the crack set with the smallest serration density at each moment is determined to be a common crack set, and the candidate tip point with the same two-dimensional coordinates at each moment is taken as the candidate tip point of the common crack set, and the candidate tip point is taken as the crack tip point.

[0133] It should be noted that under tensile force, when a crack tip appears, the crack begins to appear from that tip. As time progresses, the crack gradually expands. Therefore, by checking if the two-dimensional coordinates of the candidate crack tips at each moment are the same, we can determine that the starting point of the crack tip is the same at each moment. However, the ending point of the crack set (i.e., one of the endpoints among the candidate tips) changes continuously over time (i.e., under tensile force, as the crack expands, the ending point of the crack set changes continuously, and thus the two-dimensional coordinates change continuously). Therefore, only the true crack initiation point does not change over time, and thus its two-dimensional coordinates are different at each moment. These different two-dimensional coordinates are the ending point of the crack set. Figure 8 As shown;

[0134] When a steel component is subjected to tension, cracks will propagate at each moment. Simultaneously, new cracks may appear in the steel component over time. If a new crack tip appears at the final moment, it will not be recorded. This is because, in skeleton analysis, only when a non-zero pixel appears in its eight-neighborhood will it be recorded as a candidate tip, potentially a crack tip. When the tension stops at the final moment, even if a new crack tip appears, there is no crack propagation at the next moment, so it will not affect the fatigue of the steel component and will not be extracted in the skeleton analysis.

[0135] S26433: The crack set with the minimum serration density is divided into meshes according to the stress gradient algorithm, and the mesh stress field at each moment is captured according to the crack tip point. The stress intensity factor is calculated through the stress field.

[0136] It should be noted that by meshing the crack ensemble using the stress gradient algorithm and capturing the stress field at the crack tip, the stress state at the crack tip can be identified and analyzed more accurately. The crack tip is a critical region for material failure; accurately capturing the stress field at this point helps improve the accuracy of the analysis and avoids erroneous judgments caused by inaccurate stress field calculations. By calculating the stress field at the crack tip in real time, the stress intensity factor can be obtained more accurately, thereby accurately assessing the crack propagation trend and failure risk. This method can improve the accuracy of crack mechanics analysis, especially the computational accuracy in complex stress environments.

[0137] Specifically, such as Figure 9 As shown, in step S26433, the crack set with the minimum serration density is meshed according to the stress gradient algorithm, the mesh stress field at each moment is captured based on the crack tip point, and the stress intensity factor is calculated through the stress field. The specific operation steps are as follows:

[0138] S264331: The Euclidean distance between the crack tip and the end point of the crack set with the minimum serration density is calculated using the stress gradient algorithm. The minimum Euclidean distance between the crack tip and the end point of the crack set is selected as the radius of the crack tip.

[0139] It should be noted that when calculating the Euclidean distance between the crack tip and all end points of the crack set with the minimum serration density, the smallest Euclidean distance is selected as the radius of the crack tip. This can determine the area where the cracks are most concentrated in the crack set with the minimum serration density. Although serration density can determine the concentration point of the crack, it determines the concentration point in the entire crack region. However, the radius of the crack tip determines the concentration point of the crack, which is a more specific concentration point of the crack in the crack set with the minimum serration density (that is, as the crack changes over time, it will continue to spread outward, so the further it spreads outward, the more dispersed it becomes, and the more dispersed the crack set is, the more end points it will have).

[0140] Therefore, the radius is selected as the end point with the smallest distance from the crack tip, reflecting the most concentrated crack location in the crack set (that is, the place where stress concentration is most obvious). During crack propagation, the crack tip is the "starting point" of the crack, and the surrounding stress field and serration density are the most critical. As the crack propagates, the position of the end point will change accordingly. Therefore, the smallest Euclidean distance can help determine the most significant area of ​​crack concentration.

[0141] S264332: Divide the grid based on the radius of the crack tip;

[0142] For each grid, the stress field of the step size in the serration density is captured at each moment, starting from the time of the crack tip point.

[0143] It should be noted that in the process of calculating crack propagation, especially for complex crack morphologies, the stress field of the crack is not a simple static structure, but rather changes dynamically with time and crack propagation. Therefore, mesh generation is used to accurately capture the subtle changes in the crack propagation process, especially the stress field changes around the crack tip.

[0144] By dividing the crack region into multiple smaller regions (grids) based on the radius of the crack tip, the aim is to ensure that the stress field at and around the crack tip receives sufficient attention, because the stress concentration region at the crack tip has a decisive influence on the crack propagation rate and mode during crack propagation.

[0145] A stress field describes the stress distribution around a crack, reflecting the mechanical forces acting on the crack tip. It is not merely a simple stress value, but a complex field encompassing spatial and temporal variations. By performing stress calculations within a mesh, the stress state of different mesh elements (i.e., different locations around the crack tip) at each moment can be obtained. Cracks propagate differently at different times, and the stress field changes as the crack develops. By tracking different time steps (i.e., the crack propagation length within the radius of the crack tip can be determined based on the sawtooth density, thus utilizing sawtooth density to capture the stress field), a better understanding of the dynamic process of crack propagation can be achieved.

[0146] S264333: Calculate the stress gradient of each grid along the direction of the end point of the crack set with the minimum Euclidean distance;

[0147] The effective far-field stress (i.e., average stress) is calculated by selecting the grid with the smallest stress gradient.

[0148] The steel component body data is used as a geometric correction factor, and the effective far-field stress and the radius of the crack tip point are calculated using the geometric correction factor.

[0149] It should be noted that, firstly, the stress gradient of each grid is calculated along the direction of the crack tip (i.e., the stress gradient is the ratio of stress change to spatial distance, where spatial distance is the radius of the crack tip). This is because stress often changes drastically near the crack tip, resulting in a larger stress gradient near the crack tip. Calculating the stress gradient helps us understand the behavior of crack propagation and the stress distribution in each grid during crack tip propagation. Calculating the stress gradient can identify the stress change region near the crack tip, and selecting the grid with the smallest stress gradient can usually capture the details of crack propagation, which is helpful for stress identification.

[0150] The grid with the smallest stress gradient is selected, and the effective far-field stress (i.e., average stress) is calculated based on this grid. The purpose of this step is to find the region where the stress change is most stable in the crack propagation direction, so as to more accurately assess the stress state at the crack tip. Selecting the grid with the smallest stress gradient can avoid calculations in regions with drastic stress changes, thus avoiding local errors caused by high stress gradients. By calculating the effective far-field stress, the overall stress state near the crack tip can be obtained, which is helpful for the subsequent calculation of the stress intensity factor.

[0151] In crack analysis, the geometry correction factor is used to consider the relationship between the crack and the structural geometry, such as the crack shape, crack location, and other geometric features. Different geometries affect the stress distribution, especially near the crack tip. Therefore, the geometry correction factor is necessary to correct the calculation of the stress intensity factor according to the actual geometry of the steel member.

[0152] Geometric correction factors are typically related to the geometry of the crack (such as crack depth, length, and surface characteristics) and the material properties of the steel component (such as Poisson's ratio and Young's modulus). In calculations, a correction factor can be obtained based on the geometric data of the steel component itself (such as crack shape and aspect ratio). Because the geometric characteristics of the crack directly affect the stress concentration at the crack tip, using geometric correction factors can make the calculation of stress intensity factors more consistent with actual working conditions and avoid errors caused by geometric factors.

[0153] Example 2

[0154] like Figure 10 As shown, this application also provides a building material quality defect detection system, including: a data acquisition module 10; an analysis module 20; and a judgment module 30;

[0155] The acquisition module 10 is used to acquire images of steel components under tensile stress at each moment, preprocess each image of steel components to obtain the image of the material to be tested at each moment and the steel component body data of the image of the steel component material;

[0156] The analysis module 20 is used to identify crack regions in the image of the material under test at each time moment, and to filter the starting edge points and ending edge points of the crack regions; to filter the inflection points of the cracks based on the starting edge points and ending edge points, to calculate the crack serration degree between cracks at each time moment through the crack inflection points, and to calculate the crack propagation change amount based on the crack serration degree between cracks at each time moment.

[0157] The judgment module 30 is used to judge the fatigue detection result of the steel component material image under tensile stress by the crack propagation change.

[0158] In summary, the method and system for detecting quality defects in building materials proposed in this invention utilizes edge detection algorithms (such as Sobel and Canny) to identify crack regions in the material image at each time step, extracting edge points of the crack contour to form an edge point sequence. By identifying the edge regions of the crack, and further filtering out local maximum points in the connected integral area as positive value points of the connected surface, and connecting these two points with a straight line, a point-line is obtained. Then, by calculating the vertical distance from all intermediate edge points to the point-line, intermediate edge points deviating from the main path, i.e., edge points with the largest vertical distance, are selected as vertical distance extrema points, reflecting the most convex or concave position that the crack may correspond to. The overlap between the positive value points of the connected surface and the vertical distance extrema points is verified as inflection points, helping to identify significant change points or potential inflection points in the crack path.

[0159] Furthermore, the coordinates of the crack inflection points at each moment are determined, and the directional angles between adjacent inflection points are calculated. The crack serration degree measures the irregularity of the crack edge; a higher serration degree indicates an irregular crack edge, which may be due to local stress concentration or material heterogeneity, and usually carries a higher risk of propagation. The crack morphology is quantified by calculating the crack serration degree using the coordinates, distance, and directional angles of the inflection points. A classification threshold is set based on the crack serration degree to classify cracks into different types. Generally speaking, cracks with a larger serration degree usually have an irregular shape, strong propagation, and high fatigue risk; while cracks with a smaller serration degree are more regular, propagate more slowly, and have a lower fatigue risk.

[0160] By selecting the crack set with the highest serration density, high fatigue risk cracks are identified, and a crack database is constructed. The crack database records information about high-risk cracks at each time step. Furthermore, a skeleton algorithm is used to select the crack set with the lowest serration density at each time step. These cracks are usually more regular, close to an ideal straight line, and conform to the linear elastic fracture mechanics (LEFM) theory. The skeleton analysis counts the number of non-zero pixels in the eight-neighbor directions of each crack set, and selects pixels with one non-zero pixel as candidates for crack tip points. This method avoids relying on complex curvature calculations and directly extracts tip information from the pixels of the crack set. Subsequently, the two-dimensional coordinates of the crack tip points at each time step are determined, and it is judged whether the two-dimensional coordinates of these candidate tip points are the same. If there are different coordinates, these points are considered to be the end points of the crack set. If the coordinates are the same, these tip points belong to the same crack set, and this point is used as the common tip point of the crack to distinguish the start and end points of the crack.

[0161] Finally, the stress gradient algorithm is used to mesh the crack set with the minimum serration density, and the stress field at the crack tip is captured at each time step. By calculating the stress intensity factor, the crack propagation trend and failure risk can be evaluated.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; those skilled in the art can modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting quality defects in building materials, characterized in that, The following steps are included: Acquire images of steel components under tensile stress at each acquisition moment, preprocess each image to obtain the image of the material to be tested at each moment and the body data of the steel component from the image of the steel component material; The crack region is identified in the image of the material under test at each time moment, and the starting edge point and ending edge point of the crack region are screened. The turning point of the crack is screened according to the starting edge point and ending edge point. The crack serration degree between cracks at each time moment is calculated through the crack turning point. The crack propagation change is calculated through the crack serration degree between cracks at each time moment. The fatigue test results of the steel component material image under tensile stress are judged by the crack propagation change. A preset crack propagation change threshold y is used to determine whether the crack propagation change is greater than or equal to the crack propagation change threshold y. If so, the steel component in the material image is determined to be at risk of fatigue. For each moment in the image of the material under test, the crack region is identified, and the starting and ending edge points of the crack region are selected. The specific operation steps are as follows: The crack region is identified in the image of the material under test at each time step using an edge detection algorithm; edge points are extracted from the contour of the crack region to form an edge point sequence at multiple time steps. The Douglas-Puk algorithm is used to randomly select an edge point from the sequence of edge points in the crack region at each time step, which is then used as the starting edge point. A sliding window is created along the contour of the crack region using the starting edge point; For each sliding window, count the number of edge points, find the edge point within the sliding window that covers 1 edge point, and calculate the straight-line distance between the starting edge point and the edge point within the sliding window that covers 1 edge point; select the point with the maximum straight-line distance as the ending edge point of the crack region's contour. The specific steps for selecting crack inflection points based on the starting and ending edge points are as follows: Connect the starting edge point and the ending edge point with a straight line to obtain a point-line; For each sliding window, select sliding windows that are parallel to the line of the point, and calculate the sum of non-zero pixels for the parallel sliding windows to obtain the area of ​​the connected integral. Find the local maximum points of the edge points in the connected integral area and use them as the maximum points of the connected surface. The contour of the cracked region is filtered to select intermediate edge points, and the perpendicular distance from each intermediate edge point to the line is calculated. The edge point with the largest perpendicular distance is selected as the extreme point of the perpendicular distance. The positive value point of the connected surface is verified by coinciding with the vertical distance extreme point. When the two coincide, it is taken as the turning point.

2. The method for detecting quality defects in building materials according to claim 1, characterized in that, The crack serration is calculated at each moment between cracks by using the crack inflection point, and the crack propagation change is calculated by using the crack serration between cracks at each moment. The specific operation steps are as follows: Calculate the crack serration degree between adjacent inflection points, classify cracks based on the crack serration degree between adjacent inflection points, and obtain different types of crack sets at each time point; calculate the serration density in each type of crack set based on the number of cracks and the number of inflection points in the crack set; calculate the crack propagation change of the crack based on the serration density at each time point.

3. The method for detecting quality defects in building materials according to claim 2, characterized in that, Calculate the crack serration degree between adjacent inflection points, and classify the cracks according to the crack serration degree between adjacent inflection points to obtain the set of different types of cracks at each time point. The specific operation steps are as follows: The coordinates of the turning point at each moment are determined, and the direction angle of the adjacent turning point is calculated using the coordinates of each turning point. The crack serration degree between the adjacent turning points at each moment is calculated based on the direction angle of the adjacent turning points. Based on the crack serration degree, a crack serration degree classification threshold range q is set, and the range of crack serration degree at each moment within the crack serration degree classification threshold range q is determined to obtain multiple types of cracks at each moment. Each type of crack at each moment is then set to obtain a set of multiple types of cracks.

4. The method for detecting quality defects in building materials according to claim 3, characterized in that, The serration density of each type of crack set is calculated by combining the number of cracks and the number of inflection points in the crack set; the crack propagation change is then calculated using the serration density at each moment. The specific operation steps are as follows: For each type of crack set at each time point, calculate the number of cracks and the number of inflection points in that type of crack set, and use the number of cracks and the number of inflection points to calculate the serration density in each type of crack set at each time point. The set of cracks with the highest serration density is selected based on the serration density and identified as high fatigue risk cracks. A crack database is constructed for the set of high fatigue risk cracks at each time step; The crack propagation change is calculated using the set of high fatigue risk cracks at each moment in the crack database.

5. The method for detecting quality defects in building materials according to claim 4, characterized in that, The crack set with the highest serration density is selected as the high fatigue risk crack by filtering based on the serration density; a crack database is constructed for the high fatigue risk crack set at each time step; the crack propagation change is calculated using the high fatigue risk crack set at each time step in the crack database. The specific operation steps are as follows: The set of cracks with the highest serration density at each moment is selected by filtering based on the serration density, and these are identified as high fatigue risk cracks. The set of high fatigue risk cracks at each moment is then stored to establish a crack database. The skeleton algorithm is used to filter the set of cracks with the smallest serration density at each time step, and the skeleton analysis is performed on the pixels to obtain candidate tip points; the crack tip points are obtained based on the candidate tip points with the same coordinates of the set of cracks with the smallest serration density at each time step. The stress intensity factor at each moment is calculated using the crack tip. The crack propagation change of the crack at each moment is calculated using the set of high fatigue risk cracks in the crack database at each moment.

6. The method for detecting quality defects in building materials according to claim 5, characterized in that, The skeleton algorithm is used to filter the set of cracks with the smallest serration density at each time step, and the skeleton analysis is performed on the pixels to obtain candidate tip points. The specific operation steps are as follows: The skeleton algorithm is used to filter out the crack set with the smallest sawtooth density at each time step; skeleton analysis is performed on the crack set with the smallest sawtooth density, and the non-zero pixels in the eight-neighbor direction are counted for each pixel in the crack set with the smallest sawtooth density. By counting the number of non-zero pixels for each pixel, all pixels with a count equal to 1 are selected as candidate tip points for the crack set with the smallest serration density at each moment.

7. The method for detecting quality defects in building materials according to claim 6, characterized in that, Crack tips are obtained by identifying candidate tips at the same coordinates of the crack set with the smallest serration density type at each time step. The stress intensity factor at each time step is then calculated using these crack tips. The specific steps are as follows: The two-dimensional coordinates of the candidate tip points of the crack set with the smallest serration density type at each time moment are determined, and it is determined whether the two-dimensional coordinates of the candidate tip points of the crack set with the smallest serration density type at each time moment are the same. If not, then the two-dimensional coordinates of the candidate tip point have a single coordinate, and the candidate tip point with all single coordinates is determined to be the end point of the crack set. If so, the crack set with the smallest serration density at each moment is determined to be a common crack set, and the candidate tip point with the same two-dimensional coordinates at each moment is taken as the candidate tip point of the common crack set, and the candidate tip point is taken as the crack tip point. The crack set with the minimum serration density is divided into meshes according to the stress gradient algorithm. The mesh stress field at each moment is captured based on the crack tip point, and the stress intensity factor is calculated through the stress field.

8. The method for detecting quality defects in building materials according to claim 7, characterized in that, The crack set with the minimum serration density is meshed according to the stress gradient algorithm. The mesh stress field at each moment is captured based on the crack tip point. The stress intensity factor is calculated using the stress field. The specific operation steps are as follows: The Euclidean distance between the crack tip and the end point of the crack set with the minimum serration density is calculated using the stress gradient algorithm. The minimum Euclidean distance between the crack tip and the end point of the crack set is selected as the radius of the crack tip. A grid is created using the radius of the crack tip as a reference. For each grid, the stress field of the step size in the serration density is captured at each moment, starting from the time of the crack tip point. The stress gradient of each grid is calculated along the direction of the end point of the crack set with the minimum Euclidean distance; The effective far-field stress is calculated by selecting the grid with the smallest stress gradient. The steel component body data is used as a geometric correction factor, and the stress intensity factor is calculated using the geometric correction factor on the effective far-field stress obtained in the minimum stress gradient grid and the radius of the crack tip.

Images