Intelligent detection method and system for erosion defects of a flood discharge tunnel

Abstract

This application discloses an intelligent detection method and system for corrosion defects in flood discharge tunnels, relating to the field of tunnel corrosion detection technology. The method includes: generating a cavitation number distribution cloud map of the entire flood discharge tunnel area based on historical operating data; marking areas with cavitation values ​​below a preset cavitation number threshold as theoretical high-risk areas; marking areas with cavitation noise intensity exceeding a preset noise threshold as measured noise active areas; spatially superimposing and comparing the theoretical high-risk areas and the measured noise active areas to determine cavitation confirmation areas, potential cavitation risk areas, and cavitation anomaly areas; outputting a surface damage index through a pre-trained cavitation damage assessment model; calculating a dynamic cavitation risk factor; and generating a structured corrosion defect detection report. This method solves the problems of low detection efficiency, limited detection range, and difficulty in accurately assessing corrosion risk in existing flood discharge tunnel corrosion detection methods.

Description

Technical Field

[0001] This application relates to the field of tunnel corrosion detection technology, specifically to an intelligent detection method and system for corrosion defects in flood discharge tunnels. Background Technology

[0002] With the continuous development of water conservancy projects, flood discharge tunnels, as an important component of water conservancy hubs, undertake critical tasks such as flood discharge and sediment removal. However, during long-term operation, flood discharge tunnels are prone to corrosion defects due to the scouring of high-speed water flow and cavitation erosion. These corrosion defects not only affect the normal operation of flood discharge tunnels but may also pose a serious threat to the safety of water conservancy hubs.

[0003] Traditional methods for detecting corrosion in flood discharge tunnels mainly rely on manual inspections and localized testing. These methods suffer from low efficiency, limited scope, and difficulty in accurately assessing corrosion risk, making it difficult to conduct comprehensive and detailed testing of the entire flood discharge tunnel. Localized testing can only obtain corrosion information at specific locations and cannot reflect the overall corrosion status of the flood discharge tunnel. Summary of the Invention

[0004] This application provides an intelligent detection method and system for corrosion defects in flood discharge tunnels, which solves the technical problems of low detection efficiency, limited detection range, and difficulty in accurately assessing corrosion risk in existing flood discharge tunnel corrosion detection methods.

[0005] The technical solution to the above-mentioned technical problems in this application is as follows:

[0006] Firstly, this application provides an intelligent detection method for corrosion defects in flood discharge tunnels, the method comprising:

[0007] Based on historical operating data of the flood discharge tunnel, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated through simulation calculation;

[0008] Based on the cavitation number distribution cloud map, areas with cavitation values ​​lower than a preset cavitation number threshold are marked as theoretically high-risk areas;

[0009] Acoustic signals during the flood discharge process are collected by an underwater acoustic sensor array deployed inside the flood discharge tunnel. The cavitation noise intensity is extracted from the signals, and areas where the cavitation noise intensity exceeds a preset noise threshold are marked as active noise areas.

[0010] The theoretical high-risk area and the measured noise active area are spatially superimposed and compared to determine the cavitation erosion confirmed area, cavitation erosion potential risk area and cavitation erosion abnormal area, and differentiated UAV inspection strategies are executed for each.

[0011] Based on the data collected by UAV inspection, a surface damage index is output through a pre-trained cavitation erosion damage assessment model.

[0012] By integrating the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index, a dynamic cavitation erosion risk factor is calculated, and a structured erosion defect detection report is generated.

[0013] Secondly, this application provides an intelligent detection system for corrosion defects in flood discharge tunnels, comprising:

[0014] The cloud map drawing module is used to generate a cloud map of the cavitation number distribution of the entire flood discharge tunnel based on historical operating data of the flood discharge tunnel and through simulation calculations.

[0015] The risk classification module is used to mark areas with cavitation values ​​lower than a preset cavitation number threshold as theoretically high-risk areas based on the cavitation number distribution cloud map.

[0016] The noise segmentation module is used to collect acoustic signals during the flood discharge process through an underwater acoustic sensor array deployed inside the flood discharge tunnel, extract the cavitation noise intensity from it, and mark the area where the cavitation noise intensity exceeds the preset noise threshold as the measured noise active area.

[0017] The inspection execution module is used to spatially superimpose and compare the theoretical high-risk area with the measured noise active area to determine the cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, and execute differentiated UAV inspection strategies for each.

[0018] The model training module is used to output the surface damage index based on the data collected by UAV inspection and through a pre-trained cavitation damage assessment model.

[0019] The report generation module is used to integrate the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index to calculate the dynamic cavitation erosion risk factor and generate a structured erosion defect detection report.

[0020] This application provides one or more technical solutions, which have at least the following technical effects or advantages:

[0021] This application provides an intelligent detection method and system for erosion defects in flood discharge tunnels. First, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated using historical operational data to understand the cavitation situation at various locations, providing a basis for subsequent risk area delineation. Second, acoustic signals are collected using an underwater acoustic sensor array, and cavitation noise intensity is extracted. Areas with noise intensity exceeding a threshold are marked as measured noise active areas, supplementing theoretical analysis from a practical monitoring perspective and making risk area judgment more accurate. Then, the theoretical high-risk areas are superimposed and compared with the measured noise active areas to determine cavitation confirmation areas, potential cavitation risk areas, and cavitation anomaly areas. Differentiated UAV inspection strategies are implemented to improve detection efficiency and avoid the blindness of traditional detection methods. Different UAV inspection strategies can effectively reduce the detection of invalid areas, balancing detection efficiency and accuracy. Furthermore, based on the data collected by UAV inspections, a pre-trained cavitation damage assessment model is used to output a surface damage index, quantifying the degree of damage to the flood discharge tunnel lining surface. Finally, by integrating cavitation values, cavitation noise intensity, and surface damage index, dynamic cavitation erosion risk factors are calculated, and a structured erosion defect detection report is generated, providing a comprehensive and accurate decision-making basis for the maintenance and management of flood discharge tunnels.

[0022] Through the above technical solution, this application realizes the transformation from qualitative judgment to quantitative assessment of cavitation defects, transforming scattered detection data into dynamic cavitation risk factors and surface damage indices for risk quantification, providing accurate data support for the operation and maintenance decisions of flood discharge tunnels, fundamentally improving the accuracy, initiative and scientific nature of cavitation defects in flood discharge tunnels, realizing proactive prediction and accurate source tracing of cavitation defects, and ensuring the safe operation of tunnels. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an intelligent detection method for corrosion defects in flood discharge tunnels provided in an embodiment of this application.

[0025] Figure 2 This is a schematic diagram of the structure of an intelligent detection system for corrosion defects in a flood discharge tunnel provided in an embodiment of this application.

[0026] The components represented by each number in the attached diagram are explained below:

[0027] Cloud map drawing module 11, risk classification module 12, noise classification module 13, inspection execution module 14, model training module 15, and report generation module 16. Detailed Implementation

[0028] This application provides an intelligent detection method and system for corrosion defects in flood discharge tunnels, which addresses the technical problems of low detection efficiency, limited detection range, and difficulty in accurately assessing corrosion risks in existing flood discharge tunnel corrosion detection methods.

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0031] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0032] Example 1, as Figure 1 As shown in the figure, this application provides an intelligent detection method for corrosion defects in flood discharge tunnels, including:

[0033] S10: Based on the historical operating data of the flood discharge tunnel, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated through simulation calculation;

[0034] In this embodiment, firstly, historical operating data of the flood discharge tunnel is collected. This historical operating data includes information about the flood discharge tunnel under different time periods and water flow conditions. Combining fluid mechanics principles and relevant mathematical models, the data is analyzed and calculated in depth to obtain the cavitation number distribution across the entire flood discharge tunnel area, which is then visually presented as a cloud map. The cavitation number distribution cloud map shows the degree of cavitation at various locations within the flood discharge tunnel, providing a foundation for subsequent risk assessment and monitoring.

[0035] Specifically, step S10 in the method includes:

[0036] Collect historical operating condition data of the flood discharge tunnel within a preset historical period, wherein the historical operating condition data includes water level data, gate opening data and flow rate data;

[0037] The historical operating condition data is processed using a box plot to obtain standard historical operating condition data;

[0038] The standard historical operating condition data are classified into standard water level dataset, standard gate opening dataset, and standard flow dataset according to the operating condition type.

[0039] The median values ​​of the standard water level dataset, standard gate opening dataset, and standard flow dataset are obtained respectively, and used as representative parameter values ​​for various working conditions.

[0040] Import the three-dimensional geometric model of the flood discharge tunnel into the computational fluid dynamics simulation platform;

[0041] Based on representative parameter values ​​for various working conditions, water flow boundary condition parameters and fluid material property parameters are set in the computational fluid dynamics simulation platform.

[0042] Perform computational fluid dynamics simulations to obtain velocity and pressure field data for the entire spillway tunnel.

[0043] Based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the entire flood discharge tunnel are calculated.

[0044] Spatial interpolation and rendering are performed on the cavitation values ​​to generate a cavitation number distribution cloud map.

[0045] In this embodiment, firstly, historical operating data of the flood discharge tunnel, including water level, gate opening, and flow rate, are collected through a preset historical period. The water level data reflects the water level height of the flood discharge tunnel at different times, the gate opening data reflects the degree to which the gate is open, and the flow rate data indicates the amount of water flowing through the tunnel.

[0046] Secondly, box plots are used to process historical operating condition data, identifying and removing outliers to obtain standard historical operating condition data. This standard historical operating condition data is then categorized by operating condition type, forming standard water level datasets, standard gate opening datasets, and standard flow datasets. The median values ​​of each dataset are then obtained as representative parameter values, reducing the impact of data fluctuations and reflecting the typical characteristics of each operating condition. These operating conditions are water level data, gate opening data, and flow data.

[0047] Specifically, a box plot is a statistical chart that displays the distribution of data, including information such as the median, upper and lower quartiles, and outliers. By processing historical operating data using box plots, outliers can be identified. These outliers may be caused by measurement errors, equipment malfunctions, or other special circumstances. After removing outliers, more accurate standard historical operating data is obtained.

[0048] Next, the three-dimensional geometric model of the flood discharge tunnel was imported into a computational fluid dynamics (CFD) simulation platform. Based on representative parameter values, flow boundary condition parameters and fluid material property parameters were set to make the simulation environment more closely resemble the actual operation of the flood discharge tunnel, thus improving the realism of the simulation results. CFD simulations were executed to acquire velocity and pressure field data. Based on these data, cavitation values ​​at various locations throughout the flood discharge tunnel were calculated, reflecting the degree of cavitation at different locations. Spatial interpolation and rendering of the cavitation values ​​were performed, transforming discrete cavitation values ​​into continuous cloud maps, making the cavitation distribution more intuitive and providing accurate and visual basis for subsequent determination of theoretically high-risk areas.

[0049] The computational fluid dynamics (CFD) simulation platform is a software tool used to simulate and analyze fluid flow phenomena. Based on numerical calculation methods, it simulates the flow state of fluids under different conditions by solving fundamental equations of fluid mechanics, such as the Navier-Stokes equations. In this application, the CFD simulation platform can accurately simulate the flow of water inside a flood discharge tunnel, thereby obtaining velocity field data and pressure field data.

[0050] Furthermore, after generating the cavitation number distribution cloud map, subsequent risk area delineation is carried out based on this map. Areas with cavitation numbers below a preset cavitation number threshold are marked as theoretically high-risk areas. The marking process is based on the intrinsic relationship between cavitation number and cavitation erosion risk. The lower the cavitation number, the greater the likelihood of cavitation erosion in that area. Therefore, marking areas below the threshold helps to identify areas that may have cavitation erosion risk in advance, providing direction for subsequent detection and maintenance work.

[0051] Specifically, based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the flood discharge tunnel are calculated, including:

[0052] Based on the velocity field data and the pressure field data, the local pressure value and local velocity value of each calculation node in the simulation grid are obtained;

[0053] The local pressure value, the local flow velocity value, the saturated vapor pressure of water, and the density of water are substituted into the cavitation number definition formula for calculation to obtain the cavitation value of each calculation node.

[0054] By summing up the cavitation values ​​of all calculation nodes, the cavitation values ​​at all locations in the entire flood discharge tunnel area are obtained.

[0055] In this embodiment, firstly, the local pressure and local velocity values ​​of each computational node in the simulation mesh are extracted from the velocity field and pressure field data. The local pressure value reflects the magnitude of the water flow pressure at a specific location in the flood discharge tunnel, while the local velocity value reflects the flow velocity of the water at that location.

[0056] Secondly, the local pressure and flow velocity values, along with the saturated vapor pressure and density of water, are substituted into the cavitation number formula for calculation. The saturated vapor pressure of water is the critical pressure at which water changes from a liquid to a gaseous state at a certain temperature, while the density of water is a fundamental physical property. Through formula calculation, the cavitation values ​​for each calculation node are obtained.

[0057] Furthermore, the cavitation number (σ) = (local pressure P - saturated vapor pressure of water Pv) / (0.5 × density of water ρ × squared local flow velocity V²). Here, P - Pv represents the ability to suppress cavitation formation. The higher the pressure, the less likely the liquid is to vaporize, and the lower the risk of cavitation erosion. 0.5ρV² represents the water flow dynamics, i.e., the kinetic energy that promotes cavitation formation. The higher the flow velocity, the more significant the local pressure drop, and the higher the risk of cavitation erosion.

[0058] Therefore, the cavitation number σ reflects the relationship between "pressure that inhibits cavitation" and "kinetic energy that promotes cavitation". The smaller the cavitation number, the greater the likelihood of cavitation occurring under the current hydraulic conditions, and the higher the risk of cavitation erosion.

[0059] Finally, the cavitation values ​​from all calculated nodes are summarized to obtain the cavitation values ​​at all locations throughout the entire spillway tunnel. These cavitation values ​​serve as the foundational data for subsequent cavitation risk assessments, the delineation of theoretically high-risk zones, and the implementation of monitoring work.

[0060] S20: Based on the cavitation number distribution cloud map, areas with cavitation values ​​lower than a preset cavitation number threshold are marked as theoretical high-risk areas;

[0061] In this embodiment, the cavitation number distribution cloud map reflects the degree of cavitation at various locations throughout the flood discharge tunnel. The preset cavitation number threshold is determined based on experimental data and actual engineering experience and is used to assess the risk of cavitation erosion. Given the current hydraulic conditions in this area, regions with cavitation values ​​below this threshold are marked as theoretically high-risk areas, where the likelihood of cavitation erosion is relatively high.

[0062] Furthermore, the process of marking theoretically high-risk areas requires comparing the cavitation values ​​at each location in the cavitation number distribution cloud map with a preset threshold. Marking theoretically high-risk areas not only helps to identify areas that may have cavitation erosion risks in advance, but also provides a basis for subsequent comparison with measured noise-active areas.

[0063] Specifically, step S20 in the method includes:

[0064] Multiple underwater acoustic sensors are arranged at predetermined positions on the bottom slab and sidewalls of the flood discharge tunnel to form an underwater acoustic sensor array;

[0065] During the flood discharge process, the underwater acoustic sensor array synchronously collects the raw acoustic signals;

[0066] The original acoustic signal is subjected to bandpass filtering to extract cavitation noise signal in a preset frequency band;

[0067] Calculate the root mean square value of the cavitation noise signal as the cavitation noise intensity;

[0068] The area where the underwater acoustic sensor's cavitation noise intensity exceeds the preset noise threshold is marked as the measured noise active area.

[0069] In this embodiment, firstly, an underwater acoustic sensor array is arranged at predetermined positions on the bottom slab and sidewalls of the flood discharge tunnel. These predetermined positions are capable of monitoring acoustic signals inside the flood discharge tunnel. During the flood discharge process, the underwater acoustic sensor array synchronously collects raw acoustic signals, including various sound information generated by water flow and cavitation phenomena inside the flood discharge tunnel.

[0070] Secondly, the acquired raw acoustic signal is subjected to bandpass filtering. Bandpass filtering removes unwanted frequency bands from the raw signal, retaining only cavitation noise signals within a preset frequency band. The preset frequency band is determined based on the characteristics of cavitation noise and research experience, thereby extracting noise signals related to cavitation phenomena.

[0071] Next, the root mean square (RMS) value of the cavitation noise signal is calculated. The RMS value reflects the energy of the cavitation noise signal and is used as a measure of the cavitation noise intensity. The greater the cavitation noise intensity, the more severe the cavitation phenomenon in the region, and the higher the probability of cavitation erosion.

[0072] For example, suppose the sampling frequency of the cavitation noise signal is 1000Hz and the sampling time is 10 seconds, resulting in 10,000 sampling points. The signal values ​​of the sampling points are squared, averaged, and finally the square root is taken to obtain the specific value of the cavitation noise intensity.

[0073] By calculating cavitation noise intensity using specific numerical values, it is possible to more accurately identify areas at risk of cavitation erosion. In practical applications, preset noise thresholds can be adjusted based on the characteristics of different flood discharge tunnels and historical data.

[0074] Finally, areas where underwater acoustic sensors with cavitation noise intensity exceeding a preset noise threshold were located were marked as areas of active noise. The preset noise threshold was used to determine whether the cavitation noise in this area was abnormally active. Marking these active noise areas supplemented the theoretical analysis from a practical monitoring perspective, making subsequent assessments of cavitation erosion risk areas more accurate and comprehensive.

[0075] For example, assuming the preset noise threshold is 50dB, when the cavitation noise intensity collected and processed by an underwater acoustic sensor is 60dB, which exceeds the preset noise threshold, the area where the underwater acoustic sensor is located is marked as a measured noise active area.

[0076] S30: Acoustic signals during the flood discharge process are collected by an underwater acoustic sensor array deployed inside the flood discharge tunnel, cavitation noise intensity is extracted from them, and areas where the cavitation noise intensity exceeds a preset noise threshold are marked as measured noise active areas.

[0077] In this embodiment, an underwater acoustic sensor array is used to collect acoustic signals, providing a direct data source for monitoring cavitation phenomena inside the flood discharge tunnel. The underwater acoustic sensors can capture acoustic signals generated by water flow, cavitation formation and collapse during flood discharge. When collecting acoustic signals, it is crucial to ensure the normal operation of the sensors and accurate signal transmission, avoiding inaccurate data due to equipment malfunction or signal interference.

[0078] Furthermore, after extracting the cavitation noise intensity, areas exceeding a preset noise threshold are marked as areas of active noise. This marking process is based on the relationship between cavitation noise and cavitation erosion risk. Higher cavitation noise intensity indicates more severe cavitation in the area, and severe cavitation is often accompanied by a higher risk of cavitation erosion. Marking areas of active noise helps staff quickly locate areas where cavitation erosion may occur.

[0079] In conjunction with the marking of theoretical high-risk areas, the marking of measured noise-active areas makes the assessment of erosion risk in flood discharge tunnels more comprehensive and accurate. Theoretical high-risk areas are marked based on cavitation number distribution cloud maps, predicting areas where cavitation erosion risk may exist; while measured noise-active areas are marked using actual acoustic signals collected, reflecting the current actual cavitation situation inside the flood discharge tunnel.

[0080] Furthermore, more in-depth detection and analysis are conducted on the marked areas of active noise. For example, other detection methods, such as underwater photography and ultrasonic testing, are used to inspect the areas and determine whether corrosion defects exist and the extent of corrosion.

[0081] S40: Spatially superimpose and compare the theoretical high-risk area with the measured noise active area to determine the cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, and execute differentiated UAV inspection strategies for each.

[0082] In this embodiment, theoretical high-risk areas and measured noise-active areas are spatially superimposed and compared to assess the corrosion risk of the flood discharge tunnel by combining theoretical analysis and actual monitoring results. Cavitation confirmation areas, potential cavitation risk areas, and abnormal cavitation areas are identified. A cavitation confirmation area indicates a high risk of cavitation based on both theoretical predictions and actual monitoring; cavitation is likely already occurring, and a high-frequency, high-precision UAV inspection strategy should be implemented. UAVs can be equipped with high-resolution cameras and advanced detection instruments to conduct comprehensive, multi-angle imaging and inspection of the area, recording the specific details of the corrosion, such as its location, extent, and depth, in order to develop targeted repair plans in a timely manner.

[0083] Although the potential cavitation risk area did not show obvious cavitation noise anomalies in actual monitoring, theoretical analysis still indicates a risk of cavitation erosion. Therefore, a regular drone inspection strategy is implemented for this area. The inspection cycle can be reasonably set based on the theoretical cavitation risk level and historical data for the area. During the inspection, the focus is on monitoring changes within the area to promptly identify potential corrosion hazards.

[0084] For areas with abnormal cavitation, an emergency drone inspection strategy must be implemented immediately. Drones should quickly travel to the area to conduct in-depth inspections, analyze the causes of the anomalies, and determine whether special water flow conditions or other factors have increased the risk of cavitation. Simultaneously, continuous monitoring and observation of the area should be conducted to ensure timely understanding of its dynamic changes.

[0085] By implementing differentiated drone inspection strategies for cavitation confirmation areas, potential cavitation risk areas, and abnormal cavitation areas, the efficiency and accuracy of cavitation defect detection in flood discharge tunnels can be improved, potential problems can be identified and addressed in a timely manner, and the safe and stable operation of flood discharge tunnels can be ensured.

[0086] Specifically, step S40 in the method includes:

[0087] The theoretical high-risk area and the measured noise active area are superimposed and analyzed in the three-dimensional geometric model of the flood discharge tunnel;

[0088] The area where the theoretical high-risk zone overlaps with the measured noise active zone is defined as the cavitation confirmation zone;

[0089] The area within the theoretical high-risk zone that does not contain the measured noise active zone is defined as the potential cavitation risk zone.

[0090] The area in the measured noise active area that does not contain the theoretical high-risk area is defined as the cavitation anomaly area.

[0091] Calculate the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone, respectively.

[0092] The preset baseline number of drones is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual number of drones in each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area.

[0093] The preset baseline scanning area is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual scanning area of ​​each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area.

[0094] Based on the actual number of drones and the actual scanning area, drone inspections are performed on the cavitation confirmation area, the cavitation potential risk area, and the cavitation anomaly area, respectively.

[0095] In this embodiment, firstly, a superposition analysis of theoretical high-risk areas and measured noise-active areas is performed on the three-dimensional geometric model of the flood discharge tunnel. Through superposition, areas where both theoretical predictions and actual monitoring point to cavitation risk are identified, namely cavitation confirmation areas, indicating that the cavitation problem in these areas is already quite serious.

[0096] Secondly, the portion of the theoretically high-risk area that does not include the measured noise-active area is defined as the potential cavitation erosion risk area. Although this area did not show obvious cavitation noise anomalies in actual monitoring, theoretically, cavitation erosion is still possible. This may be due to limitations in actual monitoring, or the cavitation erosion phenomenon in this area may still be in its early stages and has not yet generated obvious noise signals. Regular monitoring and inspection of this area are necessary.

[0097] Furthermore, areas within the measured noise-active zone that do not include the theoretically high-risk zone are defined as cavitation anomaly zones. Localized abnormal water flow conditions and the special design of tunnel structures can cause discrepancies between actual monitored cavitation phenomena and theoretical predictions. For cavitation anomaly zones, emergency measures are implemented for in-depth testing to determine the cause of the anomaly.

[0098] Furthermore, after identifying the confirmed cavitation erosion area, the potential cavitation erosion risk area, and the abnormal cavitation erosion area, resource allocation parameters are calculated for each of the three areas. The calculation of resource allocation parameters requires consideration of multiple factors, such as the area size, the degree of cavitation erosion risk, and previous monitoring data. By calculating these resource allocation parameters, it is ensured that subsequent UAV inspection work can be carried out efficiently and accurately.

[0099] Then, the preset baseline number of drones is multiplied by the resource configuration parameters for each region and rounded to obtain the actual number of drones for each region. Drone resources are allocated according to the actual needs of different regions to avoid resource waste or shortage. Simultaneously, the preset baseline scanning area is multiplied by the resource configuration parameters for each region and rounded to obtain the actual scanning area for each region.

[0100] For example, assuming there are 10 baseline drones, a baseline scanning area of ​​100 square meters, a resource configuration parameter of 2 for the cavitation erosion confirmation area, a resource configuration parameter of 1 for the cavitation erosion potential risk area, and a resource configuration parameter of 3 for the cavitation erosion anomaly area.

[0101] The actual number of drones in the cavitation erosion confirmation area was 10×2=20, and the actual scanning area was 100×2=200 square meters;

[0102] The actual number of drones in the potential cavitation risk area is 10×1=10, and the actual scanning area is 100×1=100 square meters;

[0103] The actual number of drones in the cavitation anomaly area is 10×3=30, and the actual scanning area is 100×3=300 square meters.

[0104] Finally, based on the calculated actual number of drones and the actual scanned area, drone inspections were conducted in the cavitation confirmation area, the potential cavitation risk area, and the cavitation anomaly area. For the cavitation confirmation area, drones should conduct high-frequency, high-precision inspections to promptly detect changes in cavitation damage. For the potential cavitation risk area, inspections should be conducted according to a predetermined cycle, closely monitoring the dynamics within the area. For the cavitation anomaly area, drones should respond rapidly, conducting in-depth detection and continuous monitoring to promptly grasp changes in the area and provide strong support for the safe operation of the flood discharge tunnel.

[0105] The resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone are calculated respectively, including:

[0106] Obtain the lowest cavitation value and average cavitation noise intensity within the cavitation confirmation zone;

[0107] The ratio of the minimum cavitation value to the preset cavitation number threshold is calculated to obtain the first theoretical risk coefficient;

[0108] The ratio of the average cavitation noise intensity to the preset noise threshold is calculated to obtain the first measured activity coefficient;

[0109] Multiply the first theoretical risk coefficient by the first measured activity coefficient to obtain the resource allocation parameters of the cavitation confirmation zone;

[0110] Calculate the ratio of the total area of ​​the potential cavitation risk zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the potential cavitation risk zone;

[0111] Calculate the ratio of the total area of ​​the cavitation anomaly zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the cavitation anomaly zone.

[0112] In this embodiment, resource allocation parameters for the cavitation erosion confirmation zone are calculated. First, the lowest cavitation value and average cavitation noise intensity within the cavitation erosion confirmation zone are obtained. The lowest cavitation value reflects the theoretical cavitation erosion risk level of the area, while the average cavitation noise intensity reflects the actual monitored cavitation activity level. The ratio of the lowest cavitation value to a preset cavitation number threshold is calculated to obtain a first theoretical risk coefficient. The larger the first theoretical risk coefficient, the higher the theoretical cavitation erosion risk in the area. The ratio of the average cavitation noise intensity to a preset noise threshold is calculated to obtain a first measured activity coefficient. The larger this coefficient, the more severe the actual cavitation phenomenon in the area. The resource allocation parameters for the cavitation erosion confirmation zone are obtained by multiplying the first theoretical risk coefficient and the first measured activity coefficient.

[0113] For example, assuming a preset cavitation number threshold of 0.2, a preset noise threshold of 50dB, a minimum cavitation value of 0.1 within the cavitation erosion confirmation zone, and an average cavitation noise intensity of 60dB, then the first theoretical risk coefficient is 0.1÷0.2=0.5, the first measured activity coefficient is 60÷50=1.2, and the resource configuration parameter for the cavitation erosion confirmation zone is 0.5×1.2=0.6.

[0114] Secondly, the resource allocation parameters for the potential cavitation risk zone are calculated. The ratio of the total area of ​​the potential cavitation risk zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel is calculated, and then one is added to obtain the resource allocation parameters for the potential cavitation risk zone. This ratio reflects the proportion of the potential cavitation risk zone within the theoretical high-risk zone; adding one ensures that the resource allocation parameter is greater than 1, guaranteeing that the area receives sufficient inspection resources.

[0115] Finally, the resource allocation parameters for the cavitation anomaly zone are calculated. Similarly, the ratio of the total area of ​​the cavitation anomaly zone to the total area of ​​the theoretically high-risk zone of the flood discharge tunnel is calculated, and then one is added to obtain the resource allocation parameters for the cavitation anomaly zone. Because the situation in the cavitation anomaly zone is quite special, the actual cavitation phenomena monitored do not match theoretical predictions, requiring more resources for detection. By calculating this ratio and adding one, the inspection resources for this area are reasonably increased according to its area proportion.

[0116] For example, suppose the total area of ​​the theoretical high-risk zone of the flood discharge tunnel is 1000 square meters, the total area of ​​the potential cavitation risk zone is 200 square meters, and the total area of ​​the abnormal cavitation zone is 300 square meters. Then the resource allocation parameter of the potential cavitation risk zone is 200 ÷ 1000 + 1 = 1.2, and the resource allocation parameter of the abnormal cavitation zone is 300 ÷ 1000 + 1 = 1.3.

[0117] By calculating the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone, and by allocating the number of drones and the scanning area, it is possible to achieve accurate detection and efficient inspection of cavitation defects in flood discharge tunnels, promptly identify and address potential problems, and ensure the safety and stability of flood discharge tunnels during flood discharge.

[0118] S50: Based on data collected by UAV inspection, the surface damage index is output through a pre-trained cavitation damage assessment model.

[0119] In this embodiment, a pre-trained cavitation damage assessment model is used to process data collected by UAV inspections. This model can identify and analyze the erosion conditions on the surface of the flood discharge tunnel, and perform multi-dimensional analysis on the collected images, videos, and other data, including features such as the shape, size, and distribution density of the erosion.

[0120] First, the data is preprocessed to remove noise and useless information, improving data quality. Then, a feature extraction algorithm is used to extract feature vectors related to cavitation damage from the preprocessed data. These feature vectors are then input into a pre-trained cavitation damage assessment model.

[0121] Secondly, based on the input feature vector and its own training parameters and algorithm, the damage index of the flood discharge tunnel surface is calculated. The surface damage index reflects the severity of cavitation damage on the surface of the flood discharge tunnel. The higher the damage index, the more severe the corrosion on the surface of the flood discharge tunnel, and the more timely repair measures are needed.

[0122] Specifically, step S50 in the method includes:

[0123] The drone was equipped with a lidar sensor and a multispectral camera to collect three-dimensional point cloud data and multispectral image data of the lining surface of the flood discharge tunnel.

[0124] The three-dimensional point cloud data is fused with the multispectral image data to construct a three-dimensional model of the lining surface of the flood discharge tunnel.

[0125] Based on the three-dimensional model of the lining surface, the cavitation damage features of the concrete surface are identified and extracted, wherein the cavitation damage features include pitted holes, honeycomb holes and gully-like craters.

[0126] The morphological parameters of the cavitation damage characteristics are quantified using a three-dimensional morphological analysis algorithm, wherein the morphological parameters include distribution density, average depth, and total volume.

[0127] The morphological parameters are input into a pre-trained cavitation damage assessment model, which outputs a surface damage index.

[0128] In this embodiment, data is first collected using a lidar sensor and a multispectral camera mounted on a drone. The lidar sensor acquires three-dimensional point cloud data of the lining surface of the flood discharge tunnel, reflecting the geometric shape and spatial location information of the lining surface. The multispectral camera captures spectral information of different bands of the lining surface, providing data support for subsequent analysis.

[0129] Secondly, the collected 3D point cloud data is fused with multispectral image data. This process requires the use of data fusion algorithms to match and integrate the two different types of data to construct a 3D model of the lining surface of the flood discharge tunnel. The 3D model not only includes the geometry of the lining surface but also incorporates spectral information, enabling a more comprehensive and accurate representation of the actual condition of the lining surface.

[0130] Then, based on the constructed 3D model of the lining surface, the cavitation damage characteristics of the concrete surface were identified and extracted. Pitting, honeycomb, and gully-like craters are common forms of cavitation damage; these characteristics can be identified through analysis and processing of the 3D model. Using image processing and pattern recognition techniques, the data in the model was filtered and classified, and data matching the cavitation damage characteristics were extracted.

[0131] Subsequently, morphological parameters of cavitation damage characteristics were quantified using a three-dimensional morphological analysis algorithm. Distribution density reflects the distribution of cavitation damage on the lining surface, average depth reflects the severity of the damage, and total volume reflects the scale of the damage.

[0132] Finally, the quantified morphological parameters are input into a pre-trained cavitation damage assessment model. This model, trained and optimized with extensive data, accurately outputs a surface damage index based on the input morphological parameters. The surface damage index provides a reference for the maintenance and repair of flood discharge tunnels.

[0133] Furthermore, the process of constructing the cavitation erosion damage assessment model includes:

[0134] Based on historical data on cavitation damage in flood discharge tunnels, historical morphological parameters were collected to form a sample morphological parameter set.

[0135] Obtain the damage severity level corresponding to the historical morphological parameters, as assessed by experts, and quantify the damage severity level into a standard surface damage index to form a sample damage index set;

[0136] A cavitation damage assessment model is constructed based on a fully connected neural network architecture.

[0137] The cavitation erosion damage assessment model is trained under supervised supervision using the sample morphological parameter set and the sample damage index set until it is verified to converge, thus obtaining the trained cavitation erosion damage assessment model.

[0138] In this embodiment, firstly, historical cavitation damage case data of flood discharge tunnels are collected, including previous inspection reports, monitoring records, research literature, etc. Historical morphological parameters, such as the distribution density, average depth, and total volume of cavitation damage, are then collected and summarized to form a sample morphological parameter set. This sample morphological parameter set covers the cavitation damage characteristics of different flood discharge tunnels under different operating conditions.

[0139] Secondly, the damage severity level corresponding to the historical morphological parameters is obtained. The damage severity level is assessed by experts in the relevant field based on their experience and professional knowledge. By comprehensively considering the impact of cavitation damage on the structural safety and flood discharge capacity of the spillway tunnel, a damage severity level is given. Then, the damage severity level is quantified into a standard surface damage index, forming a sample damage index set. The quantification process must ensure a clear correspondence between the damage severity and the surface damage index.

[0140] Secondly, a cavitation damage assessment model is constructed based on a fully connected neural network architecture. Fully connected neural networks possess nonlinear mapping capabilities, enabling them to handle complex input-output relationships. When constructing the model, parameters such as the number of network layers and the number of neurons in each layer need to be determined.

[0141] Finally, the cavitation erosion damage assessment model was trained under supervised supervision using a set of sample morphological parameters and a set of sample damage indices. During training, the sample morphological parameters were used as input, and the corresponding sample damage indices were used as the expected output. By continuously adjusting the model's weights and biases, the model's output was made as close as possible to the expected output. Simultaneously, the model was validated using a validation set. Training was stopped when the validation error reached the convergence condition, resulting in the trained cavitation erosion damage assessment model. The trained model can accurately output the surface damage index based on the input morphological parameters, providing a reliable tool for assessing cavitation erosion damage in flood discharge tunnels.

[0142] For example, a cavitation erosion damage assessment model is built and trained based on a neural network. The specific steps are as follows:

[0143] First, data preparation involves collecting historical data on cavitation damage in flood discharge tunnels to obtain a set of sample morphological parameters.

[0144] Secondly, in model construction, the number of nodes in the input layer is equal to the dimension of the input features. For example, if there are three features in cavitation damage, namely the distribution density, average depth, and total volume, then the input layer contains three nodes. Set 1-3 hidden layers, and adjust the number of nodes in each layer through experiments, such as 64 or 32. The activation function is ReLU. The output layer generally does not use an activation function. For example, if the output takes two nodes, directly output continuous values.

[0145] Next, the model is trained, and the evaluated surface damage index is used as the output. The sample damage index set is used as the supervision label. The Adam optimizer and mean squared error loss function are used to construct the training framework. The batch size is set to 32 and the total number of training rounds is 50. An early stopping mechanism with patience=5 is introduced. When the validation set loss does not decrease for 5 consecutive rounds, the training process is automatically terminated, and the trained cavitation erosion damage assessment model is obtained.

[0146] S60: By integrating the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index, a dynamic cavitation erosion risk factor is calculated, and a structured erosion defect detection report is generated.

[0147] In this embodiment, the cavitation values, cavitation noise intensity, and surface damage index from the cavitation number distribution cloud map are integrated. The cavitation value reflects the probability of cavitation occurring inside the spillway tunnel, the cavitation noise intensity reflects the activity level of cavitation, and the surface damage index visually demonstrates the severity of cavitation erosion damage on the surface of the spillway tunnel. First, different weights are assigned to the three indicators, and the determination of these weights requires comprehensive consideration of the impact of each indicator on the erosion risk of the spillway tunnel. For example, if the surface damage index has a greater impact on the erosion risk, it is given a higher weight; if the cavitation value is theoretically more critical for predicting cavitation erosion, its weight is increased accordingly.

[0148] A dynamic cavitation risk factor is obtained by combining cavitation values, cavitation noise intensity, and surface damage index through weighted summation. This dynamic cavitation risk factor reflects the degree of cavitation risk to the flood discharge tunnel under current operating conditions. A higher risk factor indicates a greater cavitation risk to the flood discharge tunnel, requiring closer monitoring and appropriate protective measures.

[0149] Furthermore, after obtaining the dynamic cavitation risk factor, a structured cavitation defect detection report is generated using this factor as the core. The report should include basic information about the spillway tunnel, such as its location, dimensions, and design parameters; information on the UAV inspection, including the inspection area, inspection time, and data types collected; the results of the cavitation damage assessment, such as the surface damage index, characteristics, and distribution of cavitation damage; the calculation process and results of the dynamic cavitation risk factor; and targeted recommendations based on the risk factor, such as whether repair is needed, what repair measures to take, and the frequency and methods of subsequent monitoring.

[0150] Structured corrosion defect detection reports can provide comprehensive and clear information for the management and maintenance personnel of flood discharge tunnels, helping them to accurately assess the corrosion status and risk level of the flood discharge tunnels, formulate timely and scientifically sound maintenance plans, and ensure the safe and stable operation of the flood discharge tunnels.

[0151] Specifically, step S60 in the method includes:

[0152] For each region among the cavitation confirmation zone, cavitation potential risk zone, and cavitation anomaly zone, the lowest cavitation value in the cavitation number distribution cloud map, as well as the corresponding average cavitation noise intensity and surface damage index, are obtained respectively.

[0153] The initial risk value is obtained by multiplying the reciprocal of the minimum cavitation value, the average cavitation noise intensity, and the surface damage index.

[0154] The measured compressive and tensile strengths of the lining material of the flood discharge tunnel were collected.

[0155] The reference compressive strength and reference tensile strength are determined based on the design strength grade of the lining material;

[0156] Based on the measured compressive strength, measured tensile strength, reference compressive strength, and reference tensile strength, the cavitation erosion resistance coefficient of the material is calculated;

[0157] The material correction factor is calculated by dividing 1 by the cavitation resistance coefficient of the material.

[0158] Multiplying the preliminary risk value by the material correction factor yields the dynamic cavitation risk factor;

[0159] A structured corrosion defect detection report is generated, wherein the structured corrosion defect detection report includes spatial distribution information of cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, as well as corresponding dynamic cavitation risk factors.

[0160] In this embodiment, firstly, for three regions—the cavitation confirmation zone, the potential cavitation risk zone, and the cavitation anomaly zone—the lowest cavitation value, the corresponding average cavitation noise intensity, and the surface damage index are obtained from the cavitation number distribution cloud map. Since the cavitation phenomena and erosion conditions differ in different regions, obtaining data separately allows for a more accurate assessment of the risk in each region. Then, the reciprocal of the lowest cavitation value, the average cavitation noise intensity, and the surface damage index are multiplied to obtain a preliminary risk value. This preliminary risk value comprehensively considers the probability and activity level of cavitation phenomena, as well as the severity of surface damage.

[0161] Secondly, the measured compressive and tensile strengths of the lining material of the flood discharge tunnel were collected. The compressive and tensile strengths of the lining material are indicators of its resistance to cavitation erosion. Based on the design strength grade of the lining material, the benchmark compressive and tensile strengths were determined. By comparing the measured strengths with the benchmark strengths, the gap between the actual material performance and the design requirements was assessed.

[0162] Then, based on the measured compressive strength, measured tensile strength, and benchmark compressive strength and benchmark tensile strength, the values ​​are divided respectively, and then weighted and summed to calculate the material cavitation resistance coefficient. This coefficient reflects the lining material's ability to resist cavitation erosion under actual working conditions. Dividing 1 by the material cavitation resistance coefficient yields the material correction coefficient. The material correction coefficient is used to correct the initial risk value to reflect the actual cavitation risk of the flood discharge tunnel after considering material properties.

[0163] The dynamic cavitation risk factor is obtained by multiplying the initial risk value by the material correction factor. The dynamic cavitation risk factor comprehensively considers multiple factors such as cavitation phenomena, surface damage, and material properties, and can more comprehensively and accurately assess the cavitation risk of flood discharge tunnels.

[0164] For example, for a confirmed cavitation zone, assuming the lowest cavitation value in the obtained cavitation number distribution cloud map is 0.2, the corresponding average cavitation noise intensity is 80dB, and the surface damage index is 0.6, the initial risk value is 1 / 0.2 × 80 × 0.6 = 240. The measured compressive strength of the lining material in this area is collected as 35MPa, and the measured tensile strength is 3MPa. Based on its design strength grade, the baseline compressive strength is determined to be 40MPa, and the baseline tensile strength as 3.5MPa. The material cavitation resistance coefficient is calculated by weighted summation. Assuming the compressive strength weight is 0.6 and the tensile strength weight is 0.4, the material cavitation resistance coefficient is (35 / 40 × 0.6 + 3 / 3.5 × 0.4) ≈ 0.83. The material correction coefficient is 1 / 0.83 ≈ 1.2. Therefore, the dynamic cavitation risk factor for this area is 240 × 1.2 = 288.

[0165] Finally, a structured corrosion defect detection report is generated. The report includes spatial distribution information of confirmed cavitation areas, potential cavitation risk areas, and abnormal cavitation areas, as well as corresponding dynamic cavitation risk factors. The spatial distribution information provides a clear understanding of the location and extent of different risk areas within the spillway tunnel; the dynamic cavitation risk factors offer management and maintenance personnel quantifiable risk assessment results, helping to develop reasonable maintenance and repair plans based on the risk level, ensuring the safe and stable operation of the spillway tunnel.

[0166] In summary, compared with existing technologies, this application achieves intelligent detection of corrosion defects in flood discharge tunnels through steps such as collecting data with sensors and cameras mounted on UAVs, data fusion to build models, identifying and extracting damage features, quantifying morphological parameters, building and training an evaluation model, calculating dynamic cavitation risk factors by fusing multiple indicators, and generating detection reports.

[0167] In summary, the embodiments of this application have at least the following technical effects:

[0168] This application provides an intelligent detection method for erosion defects in flood discharge tunnels. First, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated using historical operational data to understand the cavitation situation at various locations, providing a basis for subsequent risk area delineation. Second, acoustic signals are collected using an underwater acoustic sensor array, and cavitation noise intensity is extracted. Areas with noise intensity exceeding a threshold are marked as measured noise active areas, supplementing theoretical analysis from a practical monitoring perspective and making risk area judgment more accurate. Then, the theoretical high-risk areas are superimposed and compared with the measured noise active areas to determine cavitation confirmation areas, potential cavitation risk areas, and cavitation anomaly areas. Differentiated UAV inspection strategies are implemented to improve detection efficiency and avoid the blindness of traditional detection methods. Different UAV inspection strategies can effectively reduce the detection of invalid areas, balancing detection efficiency and accuracy. Furthermore, based on the data collected by UAV inspections, a pre-trained cavitation damage assessment model is used to output a surface damage index, quantifying the degree of damage to the flood discharge tunnel lining surface. Finally, by integrating cavitation values, cavitation noise intensity, and surface damage index, dynamic cavitation risk factors are calculated, and a structured cavitation defect detection report is generated, providing a comprehensive and accurate decision-making basis for the maintenance and management of flood discharge tunnels. Through the above technical solution, this application achieves a transformation from qualitative judgment to quantitative assessment of cavitation defects, converting scattered detection data into dynamic cavitation risk factors and surface damage indices that quantify risk. This provides precise data support for the operation and maintenance decisions of flood discharge tunnels, fundamentally improving the accuracy, proactiveness, and scientific rigor of cavitation defects in flood discharge tunnels, enabling proactive prediction and precise source tracing of cavitation defects, and ensuring the safe operation of the tunnels.

[0169] Example 2, as Figure 2 As shown, based on the same inventive concept as the intelligent detection method for corrosion defects in flood discharge tunnels provided in Embodiment 1, this application also provides an intelligent detection system for corrosion defects in flood discharge tunnels, including:

[0170] The cloud map drawing module 11 is used to generate a cloud map of the cavitation number distribution of the entire flood discharge tunnel based on historical operating data of the flood discharge tunnel and through simulation calculation.

[0171] Risk classification module 12 is used to mark areas with cavitation values ​​lower than a preset cavitation number threshold as theoretical high-risk areas based on the cavitation number distribution cloud map;

[0172] The noise segmentation module 13 is used to collect acoustic signals during the flood discharge process through an underwater acoustic sensor array deployed inside the flood discharge tunnel, extract cavitation noise intensity from it, and mark the area where the cavitation noise intensity exceeds the preset noise threshold as the measured noise active area.

[0173] The inspection execution module 14 is used to spatially superimpose and compare the theoretical high-risk area with the measured noise active area to determine the cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, and execute differentiated UAV inspection strategies respectively.

[0174] Model training module 15 is used to output the surface damage index based on the data collected by UAV inspection and through a pre-trained cavitation damage assessment model.

[0175] The report generation module 16 is used to integrate the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity and the surface damage index to calculate the dynamic cavitation erosion risk factor and generate a structured erosion defect detection report.

[0176] In one embodiment, the cloud map drawing module 11 is specifically used for:

[0177] Collect historical operating condition data of the flood discharge tunnel within a preset historical period, wherein the historical operating condition data includes water level data, gate opening data and flow rate data;

[0178] The historical operating condition data is processed using a box plot to obtain standard historical operating condition data;

[0179] The standard historical operating condition data are classified into standard water level dataset, standard gate opening dataset, and standard flow dataset according to the operating condition type.

[0180] The median values ​​of the standard water level dataset, standard gate opening dataset, and standard flow dataset are obtained respectively, and used as representative parameter values ​​for various working conditions.

[0181] Import the three-dimensional geometric model of the flood discharge tunnel into the computational fluid dynamics simulation platform;

[0182] Based on representative parameter values ​​for various working conditions, water flow boundary condition parameters and fluid material property parameters are set in the computational fluid dynamics simulation platform.

[0183] Perform computational fluid dynamics simulations to obtain velocity and pressure field data for the entire spillway tunnel.

[0184] Based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the entire flood discharge tunnel are calculated.

[0185] Spatial interpolation and rendering are performed on the cavitation values ​​to generate a cavitation number distribution cloud map.

[0186] Furthermore, in one embodiment of the application, based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the flood discharge tunnel are calculated, including:

[0187] Based on the velocity field data and the pressure field data, the local pressure value and local velocity value of each calculation node in the simulation grid are obtained;

[0188] The local pressure value, the local flow velocity value, the saturated vapor pressure of water, and the density of water are substituted into the cavitation number definition formula for calculation to obtain the cavitation value of each calculation node.

[0189] By summing up the cavitation values ​​of all calculation nodes, the cavitation values ​​at all locations in the entire flood discharge tunnel area are obtained.

[0190] In one embodiment, the risk allocation module 12 is specifically used for:

[0191] Multiple underwater acoustic sensors are arranged at predetermined positions on the bottom slab and sidewalls of the flood discharge tunnel to form an underwater acoustic sensor array;

[0192] During the flood discharge process, the underwater acoustic sensor array synchronously collects the raw acoustic signals;

[0193] The original acoustic signal is subjected to bandpass filtering to extract cavitation noise signal in a preset frequency band;

[0194] Calculate the root mean square value of the cavitation noise signal as the cavitation noise intensity;

[0195] The area where the underwater acoustic sensor's cavitation noise intensity exceeds the preset noise threshold is marked as the measured noise active area.

[0196] In one embodiment, the inspection execution module 14 is specifically used for:

[0197] The theoretical high-risk area and the measured noise active area are superimposed and analyzed in the three-dimensional geometric model of the flood discharge tunnel;

[0198] The area where the theoretical high-risk zone overlaps with the measured noise active zone is defined as the cavitation confirmation zone;

[0199] The area within the theoretical high-risk zone that does not contain the measured noise active zone is defined as the potential cavitation risk zone.

[0200] The area in the measured noise active area that does not contain the theoretical high-risk area is defined as the cavitation anomaly area.

[0201] Calculate the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone, respectively.

[0202] The preset baseline number of drones is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual number of drones in each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area.

[0203] The preset baseline scanning area is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual scanning area of ​​each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area.

[0204] Based on the actual number of drones and the actual scanning area, drone inspections are performed on the cavitation confirmation area, the cavitation potential risk area, and the cavitation anomaly area, respectively.

[0205] Furthermore, in one embodiment of the application, the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone are calculated respectively, including:

[0206] Obtain the lowest cavitation value and average cavitation noise intensity within the cavitation confirmation zone;

[0207] The ratio of the minimum cavitation value to the preset cavitation number threshold is calculated to obtain the first theoretical risk coefficient;

[0208] The ratio of the average cavitation noise intensity to the preset noise threshold is calculated to obtain the first measured activity coefficient;

[0209] Multiply the first theoretical risk coefficient by the first measured activity coefficient to obtain the resource allocation parameters of the cavitation confirmation zone;

[0210] Calculate the ratio of the total area of ​​the potential cavitation risk zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the potential cavitation risk zone;

[0211] Calculate the ratio of the total area of ​​the cavitation anomaly zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the cavitation anomaly zone.

[0212] In one embodiment, the model training module 15 is specifically used for:

[0213] The drone was equipped with a lidar sensor and a multispectral camera to collect three-dimensional point cloud data and multispectral image data of the lining surface of the flood discharge tunnel.

[0214] The three-dimensional point cloud data is fused with the multispectral image data to construct a three-dimensional model of the lining surface of the flood discharge tunnel.

[0215] Based on the three-dimensional model of the lining surface, the cavitation damage features of the concrete surface are identified and extracted, wherein the cavitation damage features include pitted holes, honeycomb holes and gully-like craters.

[0216] The morphological parameters of the cavitation damage characteristics are quantified using a three-dimensional morphological analysis algorithm, wherein the morphological parameters include distribution density, average depth, and total volume.

[0217] The morphological parameters are input into a pre-trained cavitation damage assessment model, which outputs a surface damage index.

[0218] Furthermore, in one embodiment of the application, the process of constructing the cavitation erosion damage assessment model includes:

[0219] Based on historical data on cavitation damage in flood discharge tunnels, historical morphological parameters were collected to form a sample morphological parameter set.

[0220] Obtain the damage severity level corresponding to the historical morphological parameters, as assessed by experts, and quantify the damage severity level into a standard surface damage index to form a sample damage index set;

[0221] A cavitation damage assessment model is constructed based on a fully connected neural network architecture.

[0222] The cavitation erosion damage assessment model is trained under supervised supervision using the sample morphological parameter set and the sample damage index set until it is verified to converge, thus obtaining the trained cavitation erosion damage assessment model.

[0223] In one embodiment, the report generation module 16 is specifically used for:

[0224] For each region among the cavitation confirmation zone, cavitation potential risk zone, and cavitation anomaly zone, the lowest cavitation value in the cavitation number distribution cloud map, as well as the corresponding average cavitation noise intensity and surface damage index, are obtained respectively.

[0225] The initial risk value is obtained by multiplying the reciprocal of the minimum cavitation value, the average cavitation noise intensity, and the surface damage index.

[0226] The measured compressive and tensile strengths of the lining material of the flood discharge tunnel were collected.

[0227] The reference compressive strength and reference tensile strength are determined based on the design strength grade of the lining material;

[0228] Based on the measured compressive strength, measured tensile strength, reference compressive strength, and reference tensile strength, the cavitation erosion resistance coefficient of the material is calculated;

[0229] The material correction factor is calculated by dividing 1 by the cavitation resistance coefficient of the material.

[0230] Multiplying the preliminary risk value by the material correction factor yields the dynamic cavitation risk factor;

[0231] A structured corrosion defect detection report is generated, wherein the structured corrosion defect detection report includes spatial distribution information of cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, as well as corresponding dynamic cavitation risk factors.

[0232] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0233] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0234] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An intelligent detection method for corrosion defects in flood discharge tunnels, characterized in that, The method includes: Based on historical operating data of the flood discharge tunnel, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated through simulation calculation; Based on the cavitation number distribution cloud map, areas with cavitation values ​​lower than a preset cavitation number threshold are marked as theoretically high-risk areas; Acoustic signals during the flood discharge process are collected by an underwater acoustic sensor array deployed inside the flood discharge tunnel. The cavitation noise intensity is extracted from the signals, and areas where the cavitation noise intensity exceeds a preset noise threshold are marked as active noise areas. The theoretical high-risk area and the measured noise active area are spatially superimposed and compared to determine the cavitation erosion confirmation area, cavitation erosion potential risk area and cavitation erosion anomaly area, and differentiated UAV inspection strategies are executed for each. Based on the data collected by UAV inspection, a surface damage index is output through a pre-trained cavitation erosion damage assessment model. By integrating the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index, a dynamic cavitation erosion risk factor is calculated, and a structured erosion defect detection report is generated. Among them, based on the data collected by UAV inspection, a pre-trained cavitation erosion damage assessment model outputs a surface damage index, including: The drone was equipped with a lidar sensor and a multispectral camera to collect three-dimensional point cloud data and multispectral image data of the lining surface of the flood discharge tunnel. The three-dimensional point cloud data is fused with the multispectral image data to construct a three-dimensional model of the lining surface of the flood discharge tunnel. Based on the three-dimensional model of the lining surface, the cavitation damage features of the concrete surface are identified and extracted, wherein the cavitation damage features include pitted holes, honeycomb holes and gully-like craters. The morphological parameters of the cavitation damage characteristics are quantified using a three-dimensional morphological analysis algorithm, wherein the morphological parameters include distribution density, average depth, and total volume. The morphological parameters are input into a pre-trained cavitation erosion damage assessment model, which outputs a surface damage index. The construction process of the cavitation erosion damage assessment model includes: Based on historical data on cavitation damage in flood discharge tunnels, historical morphological parameters were collected to form a sample morphological parameter set. Obtain the damage severity level corresponding to the historical morphological parameters, as assessed by experts, and quantify the damage severity level into a standard surface damage index to form a sample damage index set; A cavitation damage assessment model is constructed based on a fully connected neural network architecture. The cavitation erosion damage assessment model is trained under supervision using the sample morphological parameter set and the sample damage index set until the model is verified to converge, thus obtaining the trained cavitation erosion damage assessment model. Specifically, by integrating the cavitation values ​​from the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index, a dynamic cavitation erosion risk factor is calculated, and a structured erosion defect detection report is generated, including: For each region among the cavitation confirmation zone, cavitation potential risk zone, and cavitation anomaly zone, the lowest cavitation value in the cavitation number distribution cloud map, as well as the corresponding average cavitation noise intensity and surface damage index, are obtained respectively. The initial risk value is obtained by multiplying the reciprocal of the minimum cavitation value, the average cavitation noise intensity, and the surface damage index. The measured compressive and tensile strengths of the lining material of the flood discharge tunnel were collected. The reference compressive strength and reference tensile strength are determined based on the design strength grade of the lining material; Based on the measured compressive strength, measured tensile strength, reference compressive strength, and reference tensile strength, the cavitation erosion resistance coefficient of the material is calculated; The material correction factor is calculated by dividing 1 by the cavitation resistance coefficient of the material. Multiplying the preliminary risk value by the material correction factor yields the dynamic cavitation risk factor; A structured corrosion defect detection report is generated, wherein the structured corrosion defect detection report includes spatial distribution information of cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, as well as corresponding dynamic cavitation risk factors.

2. The intelligent detection method for corrosion defects in flood discharge tunnels according to claim 1, characterized in that, Based on historical operational data of the flood discharge tunnel, a cavitation number distribution cloud map of the entire flood discharge tunnel area is generated through simulation calculations, including: Collect historical operating condition data of the flood discharge tunnel within a preset historical period, wherein the historical operating condition data includes water level data, gate opening data and flow rate data; The historical operating condition data is processed using a box plot to obtain standard historical operating condition data; The standard historical operating condition data are classified into standard water level dataset, standard gate opening dataset, and standard flow dataset according to the operating condition type. The median values ​​of the standard water level dataset, standard gate opening dataset, and standard flow dataset are obtained respectively, and used as representative parameter values ​​for various working conditions. Import the three-dimensional geometric model of the flood discharge tunnel into the computational fluid dynamics simulation platform; Based on representative parameter values ​​for various working conditions, water flow boundary condition parameters and fluid material property parameters are set in the computational fluid dynamics simulation platform. Perform computational fluid dynamics simulations to obtain velocity and pressure field data for the entire spillway tunnel. Based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the entire flood discharge tunnel are calculated. Spatial interpolation and rendering are performed on the cavitation values ​​to generate a cavitation number distribution cloud map.

3. The intelligent detection method for corrosion defects in flood discharge tunnels according to claim 2, characterized in that, Based on the velocity field data and the pressure field data, the cavitation values ​​at various locations throughout the entire spillway tunnel are calculated, including: Based on the velocity field data and the pressure field data, the local pressure value and local velocity value of each calculation node in the simulation grid are obtained; The local pressure value, the local flow velocity value, the saturated vapor pressure of water, and the density of water are substituted into the cavitation number definition formula for calculation to obtain the cavitation value of each calculation node. By summing up the cavitation values ​​of all calculation nodes, the cavitation values ​​at all locations in the entire flood discharge tunnel area are obtained.

4. The intelligent detection method for corrosion defects in flood discharge tunnels according to claim 1, characterized in that, Acoustic signals during the flood discharge process are collected by an underwater acoustic sensor array deployed inside the flood discharge tunnel. Cavitation noise intensity is extracted from these signals, and areas where the cavitation noise intensity exceeds a preset noise threshold are marked as active noise zones, including: Multiple underwater acoustic sensors are arranged at predetermined positions on the bottom slab and sidewalls of the flood discharge tunnel to form an underwater acoustic sensor array; During the flood discharge process, the underwater acoustic sensor array synchronously collects the raw acoustic signals; The original acoustic signal is subjected to bandpass filtering to extract cavitation noise signal in a preset frequency band; Calculate the root mean square value of the cavitation noise signal as the cavitation noise intensity; The area where the underwater acoustic sensor's cavitation noise intensity exceeds the preset noise threshold is marked as the measured noise active area.

5. The intelligent detection method for corrosion defects in flood discharge tunnels according to claim 1, characterized in that, By spatially superimposing and comparing the theoretical high-risk area with the measured noise-active area, cavitation erosion confirmation area, cavitation erosion potential risk area, and cavitation erosion anomaly area are determined, and differentiated UAV inspection strategies are implemented for each, including: The theoretical high-risk area and the measured noise active area are superimposed and analyzed in the three-dimensional geometric model of the flood discharge tunnel; The area where the theoretical high-risk zone overlaps with the measured noise active zone is defined as the cavitation confirmation zone; The area within the theoretical high-risk zone that does not contain the measured noise active zone is defined as the potential cavitation risk zone. The area in the measured noise active area that does not contain the theoretical high-risk area is defined as the cavitation anomaly area; Calculate the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone, respectively. The preset baseline number of drones is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual number of drones in each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area. The preset baseline scanning area is multiplied by the resource configuration parameters of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area, and then rounded to obtain the actual scanning area of ​​each of the erosion confirmation area, the erosion potential risk area, and the erosion anomaly area. Based on the actual number of drones and the actual scanning area, drone inspections are performed on the cavitation confirmation area, the cavitation potential risk area, and the cavitation anomaly area, respectively.

6. The intelligent detection method for corrosion defects in flood discharge tunnels according to claim 5, characterized in that, Calculate the resource allocation parameters for the cavitation confirmation zone, the cavitation potential risk zone, and the cavitation anomaly zone, respectively, including: Obtain the lowest cavitation value and average cavitation noise intensity within the cavitation confirmation zone; The ratio of the minimum cavitation value to the preset cavitation number threshold is calculated to obtain the first theoretical risk coefficient; The ratio of the average cavitation noise intensity to the preset noise threshold is calculated to obtain the first measured activity coefficient; Multiply the first theoretical risk coefficient by the first measured activity coefficient to obtain the resource allocation parameters of the cavitation confirmation zone; Calculate the ratio of the total area of ​​the potential cavitation risk zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the potential cavitation risk zone; Calculate the ratio of the total area of ​​the cavitation anomaly zone to the total area of ​​the theoretical high-risk zone of the flood discharge tunnel, and add one to obtain the resource allocation parameters of the cavitation anomaly zone.

7. An intelligent detection system for corrosion defects in flood discharge tunnels, characterized in that, An intelligent detection method for implementing the erosion defects of a flood discharge tunnel as described in any one of claims 1-6 includes: The cloud map drawing module is used to generate a cloud map of the cavitation number distribution of the entire flood discharge tunnel based on historical operating data of the flood discharge tunnel and through simulation calculations. The risk classification module is used to mark areas with cavitation values ​​lower than a preset cavitation number threshold as theoretically high-risk areas based on the cavitation number distribution cloud map. The noise segmentation module is used to collect acoustic signals during the flood discharge process through an underwater acoustic sensor array deployed inside the flood discharge tunnel, extract the cavitation noise intensity from it, and mark the area where the cavitation noise intensity exceeds the preset noise threshold as the measured noise active area. The inspection execution module is used to spatially superimpose and compare the theoretical high-risk area with the measured noise active area to determine the cavitation confirmation area, cavitation potential risk area and cavitation anomaly area, and execute differentiated UAV inspection strategies for each. The model training module is used to output the surface damage index based on the data collected by UAV inspection and through a pre-trained cavitation damage assessment model. The report generation module is used to integrate the cavitation values ​​in the cavitation number distribution cloud map, the cavitation noise intensity, and the surface damage index to calculate the dynamic cavitation erosion risk factor and generate a structured erosion defect detection report.

Images