A structural damage detection system and a structural damage detection method
By constructing a digital twin model of steel structures and using neural network algorithms, steel structure damage is identified and quantified, solving the problem of inconsistent human evaluation in existing technologies. This enables unmanned, visualized, and accurate damage detection, and predicts the service life of steel structures.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CIVIL AVIATION FLIGHT UNIV OF CHINA
- Filing Date
- 2023-02-23
- Publication Date
- 2026-06-26
AI Technical Summary
While existing machine vision-based methods for identifying steel structure damage can quickly determine the location and type of damage, they require manual evaluation of the severity, leading to inconsistent detection results and affecting objectivity.
By constructing a digital twin model of the steel structure, marking damage characteristics, using depth cameras and neural network algorithms to identify damage, and combining salt spray deposition rate and structural deformation, the damage value is quantified and simulation analysis is performed to predict the safety and service life during the lifespan.
It enables unmanned and visualized damage detection, quantifies the severity of damage, improves the objectivity and accuracy of detection, and can predict the service life of steel structures.
Smart Images
Figure CN116519699B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of structural damage detection technology, specifically to a structural damage detection system and a structural damage detection method. Background Technology
[0002] During manufacturing and use, engineering structures often suffer localized damage due to process variations or loads, such as cracks, dents, and debonding. Accumulated damage can affect the structure's normal operation. To ensure proper functioning, damage testing is necessary before delivery or after a period of use. For steel structures in critical locations, regular, multiple testing sessions are required to obtain data.
[0003] Taking bridges or aircraft components as examples, bridges need to undergo testing before opening to traffic to ensure they meet design requirements. During their service life, they require regular routine inspections or irregular special inspections to promptly detect any damage to the structure. Similarly, common steel structures in aviation, such as wing skins, must have damage below a specified level before normal use; otherwise, significant safety hazards may arise. This is especially true when these steel structures are exposed to salt spray conditions near the sea, where the corrosive effects of salt spray make them even more susceptible to damage.
[0004] In recent years, with the booming development of emerging industries such as information science and artificial intelligence, computer technology has been increasingly applied to all aspects of national economic production, including industry, agriculture, and modern service industries. Machine vision technology, based on image sensing devices such as cameras, enables computers to observe, recognize, and understand the world.
[0005] Machine vision has been widely used in hazardous working environments where manual labor is unsuitable or in scenarios where human vision cannot meet measurement or inspection requirements, such as in engineering or aerospace structures. For modern large-scale industrial production, manual inspection of product quality is inefficient and lacks precision, while machine vision technology can effectively improve production efficiency and inspection accuracy, thereby shortening product production cycles and saving operating costs.
[0006] However, while existing machine vision-based methods for identifying damage to steel structures can quickly determine the location and type of damage, the severity of the damage still requires human evaluation based on experience after the damage detection is completed. This can lead to inconsistent judgments from different people, indirectly interfering with the objectivity of the detection results. Summary of the Invention
[0007] (a) Technical problems to be solved
[0008] To address the shortcomings of existing technologies, this invention provides a structural damage detection system and method. The system comprises: a modeling unit to construct a digital twin model of the steel structure; a marking unit to mark identified damage features within the digital twin model; an evaluation unit to assess the identified steel structure damage and obtain damage values; and an analysis unit to obtain damage correlation values based on the digital twin model and damage assessment, determining the impact of damage evolution on the steel structure. Based on the magnitude of the damage correlation values, the impact of damage evolution on the steel structure is quantified, and the safety of the steel structure during its lifespan is evaluated. Furthermore, through the digital twin analysis model, simulation analysis of the evolution of crack ratio Lw and deflection Ni can predict the service life of the steel structure, thus solving the problems in the prior art.
[0009] (II) Technical Solution
[0010] To achieve the above objectives, the present invention provides the following technical solution: a structural damage detection method, comprising: imaging a steel structure, constructing an image library of the steel structure, and constructing a digital twin model of the steel structure based on the images in the image library and the material properties of the steel structure; comprising: cleaning the steel structure using a cleaning device, selecting a depth imaging device to perform a circumferential scanning image along the length of the steel structure, connecting the imaging positions one after another, acquiring the steel structure image library, and marking each image position; based on the position markings, converting the steel structure images into a steel structure model using a 3D SLAM algorithm based on a depth camera; acquiring the dimensions and material properties of the steel structure, marking them in the steel structure model, and constructing a digital twin model of the steel structure;
[0011] By constructing a damage identification model, damage to the steel structure is identified, and the identified damage features are marked on the digital twin model of the steel structure. The identified steel structure damage is evaluated, the damage value of the steel structure is obtained, and it is determined whether the damage value exceeds the threshold. If it exceeds the threshold, the individual damage with the degree of damage exceeding the corresponding threshold and its corresponding location are identified and marked significantly on the digital twin model of the steel structure.
[0012] Based on the digital twin model of steel structure and damage assessment, damage correlation values are obtained, and simulation analysis of steel structure loss is performed to determine the impact of damage evolution on the steel structure and determine the safety of the steel structure during its lifespan. A steel structure maintenance mode library is established. During the lifespan, if the damage correlation value is lower than expected, structural damage with a single damage level exceeding the threshold is identified, and the corresponding maintenance method is found in the library.
[0013] Furthermore, based on deep web crawlers, deep searches are conducted on the web to obtain high-frequency damage on steel structures under salt spray conditions, and a steel structure damage database is established. A damage recognition model is established based on neural convolution algorithms. Using materials from the steel structure image database and the steel structure damage database, training and testing sets are established to complete the training and testing of the damage recognition model.
[0014] Furthermore, based on the generated damage recognition model, images in the steel structure image library are identified to obtain damage features. Images of steel structures with damage and similarity higher than the threshold are marked and output. A damage structure classifier is built based on a deep neural network algorithm. Based on the identified damage features of the steel structures, the identified steel structures are classified, and the digital twin models of the steel structures are marked accordingly.
[0015] Furthermore, based on the salt spray deposition rate in the salt spray environment of the steel structure, the average salt spray deposition rate is obtained along the length of the steel structure; several sets of average salt spray deposition rates are obtained along the time axis, and a fitting function for the change of the average salt spray deposition rate is established based on the changing trend of the average salt spray deposition rate.
[0016] After the fitting function of the average change of salt spray deposition rate is subjected to KS test, the maximum value of the average salt spray deposition rate in the next multiple time periods is obtained as the salt spray deposition rate; based on image recognition, when there are cracks on the surface of the steel structure, the sum of the ratios of the crack depth to the cross-sectional diameter of the steel structure is detected to obtain the crack ratio; under the load condition of the steel structure, the degree of vertical deformation along the vertical direction of the steel structure is obtained to determine the deflection value.
[0017] Furthermore, the salt spray deposition rate Yw, crack ratio Lw, and deflection N are obtained, dimensionless processed, and correlated to obtain the steel structure damage value Gs. The steel structure damage value Gs is compared with the corresponding threshold to determine whether the steel structure damage value Gs is greater than the threshold. If it is greater than the threshold, under the current salt spray deposition rate Yw, the portion of crack ratio Lw and deflection N that is greater than the corresponding threshold is obtained, and combined with the corresponding location, it is marked significantly on the digital twin model of the steel structure.
[0018] Furthermore, the method for obtaining the steel structure damage value Gs conforms to the following formula:
[0019]
[0020] Parameters: Salt spray factor Ay, 0.11≤Ay≤0.87, crack factor Al, 0.18≤Al≤0.53, deflection factor An, 0.81≤An≤1.65, E is a constant correction coefficient.
[0021] Furthermore, the salt spray deposition rate Yw, crack ratio Lw, and deflection N are marked on the digital twin model of the steel structure to form a digital twin analysis model of the steel structure. Based on the current salt spray deposition rate Yw, without increasing the load on the steel structure, several sets of crack ratio Lw and deflection N data are sequentially obtained from the digital twin analysis model of the steel structure along the time cycle.
[0022] The crack ratio Lw and deflection N are dimensionless and correlated to form a damage correlation value Ss(Lw, N). The simulation test cycle is adjusted until a new damage correlation value Ss(Lw, N) is obtained. It is then compared with the corresponding threshold. If it is greater than the threshold, the corresponding damage is marked and output.
[0023] Furthermore, the association method for the damage correlation value Ss(Lw, N) conforms to the following formula:
[0024]
[0025] Where 0≤k1≤1, 0≤k2≤1, and k1 2 +k2 2 =1, k2 and k1 are weights, and their specific values are adjusted and set by the user;
[0026]
[0027] Among them, Lw i N represents the expected average crack ratio. i This represents the expected mean of deflection.
[0028] Furthermore, high-frequency damage on the steel structure is obtained under salt spray conditions, and a steel structure maintenance model library is constructed based on the maintenance methods under high-frequency damage conditions; under the current salt spray settling rate Yw, the simulation test cycle is adjusted until the crack ratio Lw and deflection N are greater than the corresponding thresholds within the service life and when the damage correlation value Ss(Lw, N) is greater than the threshold.
[0029] Based on the damage identification model, damage characteristics on the steel structure are identified, and corresponding maintenance methods are selected from the steel structure maintenance model library based on the similarity of the damage characteristics.
[0030] A structural damage detection system includes: a modeling unit for imaging a steel structure and constructing a digital twin model of the steel structure based on the imaging and the material properties of the steel structure; a marking unit for identifying damage to the steel structure and marking the identified damage features in the digital twin model of the steel structure; and an evaluation unit for evaluating the identified steel structure damage, obtaining the steel structure damage value, and if the steel structure damage value exceeds a threshold, identifying the single damage exceeding the corresponding threshold and its corresponding location.
[0031] The analysis unit, based on the digital twin model of the steel structure and damage assessment, obtains damage correlation values, performs simulation analysis on the steel structure loss, and judges the impact of damage evolution on the steel structure; the retrieval unit establishes a steel structure maintenance mode library. If the damage correlation value is lower than expected, it determines the structural damage with a damage level higher than the threshold and finds the corresponding maintenance method in the library.
[0032] (III) Beneficial Effects
[0033] This invention provides a structural damage detection system and method, which have the following advantages:
[0034] By establishing a steel structure damage database and a steel structure image database, a damage recognition model is constructed to identify damage features on the steel structure. These features are then marked on the digital twin model of the steel structure. This results in a higher degree of automation during steel structure inspection. By constructing a digital twin model of the steel structure and marking the identified damage features on it, the level of visualization during steel structure damage inspection is higher, thus replacing visual inspection.
[0035] Based on the current salt spray deposition rate Yw, the magnitude of the steel structure damage value Gs is determined, thereby determining whether the steel structure can continue to be used under salt spray conditions. When damage exists, a specific evaluation of the severity of the damage can be made. After completing the detection of steel structure damage, the severity of the damage can be quickly determined, and the damage detection can be quantified.
[0036] Based on the judgment of the magnitude of the damage correlation value, the impact of damage evolution on the steel structure is quantified, the safety of the steel structure during its service life is evaluated, and the evolution of crack ratio Lw and deflection Ni is simulated and analyzed through the digital twin analysis model of the steel structure, which can predict the service life of the steel structure. Attached Figure Description
[0037] Figure 1 This is a schematic diagram of the structural damage detection method of the present invention;
[0038] Figure 2 This is a schematic diagram of the structural damage detection system of the present invention;
[0039] Figure 3 This is a schematic diagram of the damage correlation value structure of the present invention.
[0040] In the picture:
[0041] 10. Modeling unit; 20. Labeling unit; 30. Evaluation unit; 40. Analysis unit; 50. Retrieval unit. Detailed Implementation
[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. 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.
[0043] Example 1
[0044] Please see Figure 1-3 This invention provides a structural damage detection method, comprising the following steps:
[0045] Step 1: Image the steel structure and build an image library of the steel structure. Based on the images in the image library and the material properties of the steel structure, construct a digital twin model of the steel structure.
[0046] Step one includes the following:
[0047] Step 101: Clean the steel structure using cleaning equipment, such as a brush or a high-speed water gun; remove any adhering substances from the surface of the steel structure; select a depth-of-field imaging device to scan and image along the length of the steel structure, connecting the imaging positions one by one to obtain a steel structure image library, and mark each image position according to its position.
[0048] Step 102: Based on the location markers, the steel structure image is converted into a steel structure model using a 3D SLAM algorithm based on a depth camera; the dimensions and material properties of the steel structure are obtained, including at least the mechanical properties, density, and stiffness of the material; the data are then marked into the steel structure model to construct a data twin model of the steel structure.
[0049] In use, by following steps 101 to 102, a data twin model of the steel structure is constructed based on imaging the steel structure, measuring its dimensions, and acquiring its performance data. After the initial inspection of the steel structure is completed, the detected damage features can be marked on the data twin model of the steel structure.
[0050] Step 2: By constructing a damage identification model, the damage to the steel structure is identified, and the identified damage features are marked in the digital twin model of the steel structure.
[0051] Step two includes the following:
[0052] Step 201: Based on deep web crawling, perform deep search on the web to obtain high-frequency damage on steel structures under salt spray conditions and establish a steel structure damage database; the steel structure damage includes at least cracks, surface scratches, dents, broken nails, bending deformation, etc.
[0053] Step 202: Establish a damage recognition model based on the neural convolution algorithm. Use materials from the steel structure image library and the steel structure damage library (i.e., images of various types) to establish training and testing sets, and complete the training and testing of the damage recognition model.
[0054] Step 203: Based on the generated damage recognition model, identify the images in the steel structure image library, obtain damage features, and mark and output the images of steel structures that have damage and whose similarity is higher than the threshold.
[0055] Step 204: Build a damage structure classifier based on the deep neural network algorithm. Classify the identified steel structures according to the damage characteristics of the identified steel structures, and then label the digital twin models of the steel structures in turn.
[0056] In practice, combining steps 201 to 204, a damage recognition model is constructed by establishing a steel structure damage database and a steel structure image database to identify damage features on the steel structure. After classification, the digital twin models of the steel structure are marked sequentially. At this point, the initial detection of steel structure damage is completed with the help of depth imaging equipment. If the depth imaging equipment is carried by a drone or similar device, the degree of unmanned operation is higher when inspecting the steel structure. Moreover, by constructing a digital twin model of the steel structure and marking the identified damage features on the digital twin model, the visualization level is higher when detecting steel structure damage. Under certain conditions, it can replace visual inspection.
[0057] Step 3: Assess the identified steel structure damage, obtain the steel structure damage value, determine whether the damage value exceeds the threshold, and if it exceeds the threshold, identify the single damage that exceeds the corresponding threshold and its corresponding location, and mark it significantly on the steel structure digital twin model.
[0058] Step three includes the following:
[0059] Step 301: Based on the salt spray deposition rate in the salt spray environment of the steel structure, obtain the average salt spray deposition rate along the length of the steel structure.
[0060] Several sets of average salt spray deposition rates (more than ten sets) were obtained along the time axis. Based on the changing trend of the average salt spray deposition rate, a fitting function for the change of the average salt spray deposition rate was established.
[0061] After the fitting function of the average change of salt spray deposition rate is subjected to KS test, the maximum value of the average salt spray deposition rate over the next multiple time periods (at least more than ten periods) is obtained as the salt spray deposition rate.
[0062] Step 302: Based on image recognition, when there are cracks on the surface of the steel structure, detect the ratio of crack depth to the cross-sectional diameter of the steel structure to obtain the crack ratio; when there are several cracks on the surface of the steel structure, use the sum of the crack ratios of the several cracks as the crack ratio.
[0063] Under load conditions, the degree of vertical deformation along the vertical direction of the steel structure is obtained to determine the deflection value.
[0064] In use, the damage to steel structures can be characterized based on salt spray deposition rate, crack ratio, and deflection. It should be noted that there are many types of damage to steel structures, including at least cracks, surface scratches, pits, broken nails, and bending deformation. These damages are identified in steps one and two. After step three, the most representative crack ratio and deflection are selected to evaluate the degree of damage to the steel structure.
[0065] Step 303: Obtain the salt spray deposition rate Yw, crack ratio Lw, and deflection N, perform dimensionless processing, and obtain the steel structure damage value Gs after correlation.
[0066] The damage value Gs of the steel structure is obtained according to the following formula:
[0067]
[0068] The parameters are defined as follows: salt spray factor Ay, 0.11≤Ay≤0.87, crack factor Al, 0.18≤Al≤0.53, deflection factor An, 0.81≤An≤1.65, and E is a constant correction coefficient.
[0069] It should be noted that: the salt spray factor Ay, crack factor Al, and deflection factor An are determined by collecting multiple sets of sample data by those skilled in the art and setting corresponding preset proportional coefficients for each set of sample data; the preset proportional coefficients and the collected sample data are substituted into the formulas, and any three formulas form a system of three linear equations; the calculated coefficients are filtered and the average value is taken to determine the values of the salt spray factor Ay, crack factor Al, and deflection factor An.
[0070] The size of the coefficient is a specific value obtained by quantifying each parameter to facilitate subsequent comparison. The size of the coefficient depends on the amount of sample data and the preset proportional coefficient initially set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantified value.
[0071] Step 304: Compare the steel structure damage value Gs with the corresponding threshold. Based on the comparison result, determine whether the steel structure damage value Gs is greater than the threshold. If it is not greater than the threshold, the steel structure can continue to be used under the current salt spray conditions.
[0072] If the values are greater than the threshold, then under the current salt spray deposition rate Yw, the portions of crack ratio Lw and deflection N that are greater than the corresponding thresholds are obtained, and combined with the corresponding locations, they are marked significantly on the digital twin model of the steel structure.
[0073] When used, in conjunction with steps 301 to 304: based on the current salt spray deposition rate Yw, determine the magnitude of the steel structure damage value Gs, thereby determining whether the steel structure can continue to be used under salt spray conditions. In this way, when damage exists, a specific evaluation of the severity of the damage can be formed. After completing the detection of steel structure damage, the severity of the damage can be quickly determined, and the damage detection can be quantified.
[0074] Step 4: Based on the digital twin model of the steel structure and damage assessment, obtain the damage correlation value, conduct simulation analysis on the steel structure loss, determine the impact of damage evolution on the steel structure, and determine the safety of the steel structure during its service life.
[0075] Step four includes the following:
[0076] Step 401: Mark the salt spray deposition rate Yw, crack ratio Lw, and deflection N on the digital twin model of the steel structure to form a digital twin analysis model of the steel structure.
[0077] Step 402: Based on the current salt spray deposition rate Yw, without increasing the load on the steel structure, sequentially obtain several sets of crack ratio Lw and deflection N data from the digital twin analysis model of the steel structure along a time cycle (e.g., 5 days, 10 days, or other suitable testing cycles); for example, Lw n-1 Lw n Lw n+1 and N n-1 N n N n+1 wait;
[0078] The crack ratio Lw and deflection N are dimensionless and then correlated to form the damage correlation value Ss(Lw, N); the correlation method conforms to the following formula:
[0079]
[0080]
[0081] Where 0≤k1≤1, 0≤k2≤1, and k1 2 +k2 2 =1, k2 and k1 are weights, and their specific values are adjusted and set by the user;
[0082]
[0083] Among them, Lwi N represents the expected average crack ratio. i This represents the expected mean of deflection.
[0084] When in use, by determining the damage correlation value Ss(Lw, N), it is possible to monitor the changes in crack ratio Lw and deflection N under the current salt spray deposition rate Yw. If the changes exceed expectations, timely evaluation can be conducted.
[0085] Step 403: Adjust the simulation test cycle until a new damage correlation value Ss(Lw, N) is obtained. Compare it with the corresponding threshold. If it is greater than the threshold, the corresponding damage is marked and output.
[0086] In use, by acquiring damage correlation values and judging the magnitude of the damage correlation values, the impact of damage evolution on steel structures can be quantified to a certain extent, thereby evaluating the safety of steel structures during their service life. Moreover, through the digital twin analysis model of steel structures, the evolution of crack ratio Lw and deflection Ni can be simulated and analyzed, and the service life of steel structures can also be predicted.
[0087] Step 5: Establish a steel structure maintenance model library. During the service life, if the damage correlation value is lower than expected, identify structural damage with a single damage level exceeding the threshold and find the corresponding maintenance method in the library.
[0088] Step five includes:
[0089] Step 501: Under salt spray conditions, high-frequency damage on the steel structure is obtained, and a steel structure maintenance model library is constructed based on the maintenance methods under high-frequency damage conditions.
[0090] Step 502: Under the current salt spray deposition rate Yw, adjust the simulation test cycle until, within the lifespan and when the damage correlation value Ss(Lw, N) is greater than the threshold, the portion of crack ratio Lw and deflection N that is greater than the corresponding threshold.
[0091] Step 503: Based on the damage identification model, identify the damage characteristics on the steel structure, and select the corresponding maintenance method from the steel structure maintenance model library based on the similarity of the damage characteristics.
[0092] When damage exists on the steel structure, the digital twin analysis model of the steel structure is used to determine the changing trends of the crack ratio Lw and deflection N. If it is determined that the subsequent changes in the crack ratio Lw and deflection N will have a significant impact on the normal use of the steel structure, the corresponding maintenance method is output based on the similarity of damage characteristics according to the constructed steel structure maintenance model library. This allows users to quickly obtain the corresponding maintenance methods after completing the steel structure damage detection, which has a wider coverage compared to simple damage detection.
[0093] Example 2
[0094] Please see Figure 1-3 This invention provides a structural damage detection system, comprising:
[0095] Modeling Unit 10: Image the steel structure and construct a digital twin model of the steel structure based on the imaging and the material properties of the steel structure.
[0096] Marking unit 20 identifies damage to the steel structure and marks the identified damage features in the digital twin model of the steel structure;
[0097] Assessment Unit 30 assesses the identified steel structure damage, obtains the steel structure damage value, and if the steel structure damage value exceeds the threshold, identifies the single damage that exceeds the corresponding threshold and its corresponding location.
[0098] Analysis Unit 40: Based on the digital twin model of the steel structure and damage assessment, obtain damage correlation values, perform simulation analysis on the steel structure loss, and determine the impact of damage evolution on the steel structure.
[0099] Retrieval Unit 50: Establish a steel structure maintenance model library. If the damage correlation value is lower than expected, determine the structural damage with a damage level higher than the threshold, and find the corresponding maintenance method in the library.
[0100] Based on the above, at least the following effects exist in this application:
[0101] Based on the digital twin analysis model of the steel structure, the changing trends of crack ratio Lw and deflection N are determined. If it is determined that the subsequent changes in crack ratio Lw and deflection N will have a significant impact on the normal use of the steel structure, then based on the constructed steel structure maintenance model library, the corresponding maintenance methods are output through the similarity of damage characteristics. This allows for the rapid acquisition of appropriate maintenance methods, which has a wider coverage compared to simple damage detection.
[0102] By establishing a steel structure damage database and a steel structure image database, a damage recognition model is constructed to identify damage features on the steel structure. These features are then marked on the digital twin model of the steel structure. This results in a higher degree of automation during steel structure inspection. By constructing a digital twin model of the steel structure and marking the identified damage features on it, the level of visualization during steel structure damage inspection is higher, thus replacing visual inspection.
[0103] Based on the current salt spray deposition rate Yw, the magnitude of the steel structure damage value Gs is determined, thereby determining whether the steel structure can continue to be used under salt spray conditions. When damage exists, a specific evaluation of the severity of the damage can be made. After completing the detection of steel structure damage, the severity of the damage can be quickly determined, and the damage detection can be quantified.
[0104] Based on the judgment of the magnitude of the damage correlation value, the impact of damage evolution on the steel structure is quantified, the safety of the steel structure during its service life is evaluated, and the evolution of crack ratio Lw and deflection Ni is simulated and analyzed through the digital twin analysis model of the steel structure, which can predict the service life of the steel structure.
[0105] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0106] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0107] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0108] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0109] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0110] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0111] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
[0113] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for detecting structural damage, characterized in that: include: The steel structure is imaged, an image library of steel structures is constructed, and a digital twin model of the steel structure is built based on the images in the image library and the material properties of the steel structure; including: The steel structure is cleaned using cleaning equipment. A depth imaging device is then used to scan and image the steel structure along its length, with the imaging positions connected sequentially to obtain a steel structure image library. Each image position is then marked. Based on the position markings, a 3D SLAM algorithm based on a depth camera is used to convert the steel structure images into a steel structure model. The dimensions and material properties of the steel structure are then obtained and marked into the steel structure model to construct a data twin model of the steel structure. By constructing a damage identification model, damage to the steel structure is identified, and the identified damage features are marked in the digital twin model of the steel structure. The identified steel structure damage is assessed, the steel structure damage value is obtained, and it is determined whether the damage value exceeds the threshold. If it exceeds the threshold, the individual damage with the degree of damage exceeding the corresponding threshold and its corresponding location are identified and marked significantly on the digital twin model of the steel structure. Based on the digital twin model of steel structure and damage assessment, damage correlation values are obtained, and simulation analysis of steel structure loss is performed to determine the impact of damage evolution on steel structure and to determine the safety of steel structure during its lifespan. Establish a steel structure maintenance model library. During the service life, if the damage correlation value is lower than expected, identify structural damage with a single damage level exceeding the threshold and find the corresponding maintenance method in the library. Based on the salt spray deposition rate in the salt spray environment of the steel structure, the average salt spray deposition rate is obtained along the length of the steel structure; several sets of average salt spray deposition rates are obtained along the time axis, and a fitting function for the change of the average salt spray deposition rate is established based on the changing trend of the average salt spray deposition rate. After performing the KS test on the fitting function of the average change of salt spray deposition rate, the maximum value of the average salt spray deposition rate over the next multiple time periods is obtained as the salt spray deposition rate. Based on image recognition, when cracks exist on the surface of a steel structure, the sum of the ratios of crack depth to the cross-sectional diameter of the steel structure is detected to obtain the crack ratio; when the steel structure is under load, the degree of vertical deformation along the vertical direction of the steel structure is obtained to determine the deflection value. The salt spray deposition rate Yw, crack ratio Lw, and deflection N were obtained, dimensionless processing was performed, and after correlation, the steel structure damage value Gs was obtained. The steel structure damage value Gs is compared with the corresponding threshold to determine whether the steel structure damage value Gs is greater than the threshold. If it is greater than the threshold, then under the current salt spray deposition rate Yw, the crack ratio Lw and deflection N that are greater than the corresponding threshold are obtained and marked significantly on the digital twin model of the steel structure in combination with the corresponding location. The damage value Gs of the steel structure is obtained according to the following formula: Parameter: Salt spray factor , , Deflection factor , E is a constant correction factor; Salt spray deposition rate Yw, crack ratio Lw, and deflection N are marked on the digital twin model of the steel structure to form a digital twin analysis model of the steel structure. Based on the current salt spray deposition rate Yw, without increasing the load on the steel structure, several sets of crack ratio Lw and deflection N data are sequentially obtained from the digital twin analysis model of the steel structure along the time cycle. The crack ratio Lw and deflection N are made dimensionless and then correlated to form a damage correlation value. Adjust the simulation test cycle until new damage correlation values are obtained. The damage is compared with the corresponding threshold. If it is greater than the threshold, the corresponding damage is marked and output. Damage correlation value The association method conforms to the following formula: in, 1, ,and 1, This is the weight, and its specific value is adjusted and set by the user. in, This represents the expected average crack ratio. This represents the expected mean of the deflection. High-frequency damage on steel structures was acquired under salt spray conditions, and a steel structure maintenance model library was constructed based on maintenance methods under high-frequency damage conditions. Under the current salt spray deposition rate Yw, adjust the simulation test cycle until it is within the lifespan and the damage correlation value is... When the values are greater than the threshold, the portions of the crack ratio Lw and deflection N that are greater than the corresponding threshold; Based on the damage identification model, damage characteristics on the steel structure are identified, and corresponding maintenance methods are selected from the steel structure maintenance model library based on the similarity of the damage characteristics.
2. The structural damage detection method according to claim 1, characterized in that: Based on deep web crawlers, we conduct deep searches on the web to obtain high-frequency damage on steel structures under salt spray conditions and establish a steel structure damage database. A damage recognition model is established based on the neural convolution algorithm. Training and testing sets are built using materials from a steel structure image library and a steel structure damage library to complete the training and testing of the damage recognition model.
3. The structural damage detection method according to claim 2, characterized in that: Based on the generated damage recognition model, images in the steel structure image library are identified to obtain damage features. Images of steel structures with damage and similarity higher than the threshold are marked and output. A damage structure classifier was built based on a deep neural network algorithm. Based on the identified damage characteristics of the steel structures, the identified steel structures were classified, and then the digital twin models of the steel structures were labeled.
4. A structural damage detection system, characterized in that: The detection method according to any one of claims 1-3 includes: Modeling unit (10) images the steel structure and constructs a digital twin model of the steel structure based on the imaging and the material properties of the steel structure; The marking unit (20) identifies damage to the steel structure and marks the identified damage features in the digital twin model of the steel structure. The assessment unit (30) assesses the identified steel structure damage and obtains the steel structure damage value. If the steel structure damage value exceeds the threshold, it determines the single damage that exceeds the corresponding threshold and its corresponding location. Analysis unit (40) obtains damage correlation values based on the digital twin model of steel structure and damage assessment, performs simulation analysis on steel structure loss, and judges the impact of damage evolution on steel structure. The retrieval unit (50) establishes a steel structure maintenance model library. If the damage correlation value is lower than expected, it determines the structural damage with a damage level higher than the threshold and finds the corresponding maintenance method in the library.