Substation reinforced concrete corrosion thickness evaluation method and system based on ultrasonic waves

Abstract

The invention belongs to the technical field of corrosion thickness measurement, and provides a transformer substation reinforced concrete corrosion thickness evaluation method and system based on ultrasonic waves, and the method comprises the steps: determining an ultrasonic parameter combination according to geometrical shape information and spatial position information, and achieving the real-time adjustment of ultrasonic parameters; according to the frequency spectrum characteristic parameters and a preset support vector machine classification model, obtaining a corrosion type; the corrosion thickness and the corrosion degree grade are obtained according to the corrosion type and corrosion degree evaluation model, and the influence of different types and distribution characteristics of corrosion products on propagation and reflection of ultrasonic waves on the surface of the component is solved; through the comprehensive improvement of adjusting ultrasonic detection parameters to comprehensively reflect the overall corrosion condition, considering the geometrical shape and the spatial position of the corrosion part and performing evaluation according to the corrosion type, the evaluation method can adapt to the complex corrosion part, the detection data quality is improved, the comprehensive and accurate evaluation of the corrosion degree of the metal component is realized, and the evaluation efficiency is improved. And a reliable basis is provided for component maintenance and safety assessment.

Description

Technical Field

[0001] The invention belongs to the technical field of corrosion thickness measurement, and in particular relates to an ultrasonic-based method and system for assessing the corrosion thickness of reinforced concrete in a transformer substation. Background Art

[0002] Substations are important hubs in the power system, responsible for the transmission and distribution of electric energy. Once a failure occurs, it may cause a large-scale power outage, so it is crucial to ensure the safe and stable operation of substations. Limited by the background of the times and construction technology, early substations mostly used reinforced concrete structures. However, due to the lack of durability of concrete, as well as the influence of the technical level and construction quality at the time, concrete frames generally have varying degrees of aging and deterioration problems.

[0003] Among the many factors that endanger the durability of concrete structures, steel corrosion is one of the most serious factors. Steel corrosion will generate rust, resulting in a reduction in cross-sectional area and causing cracks on the surface and inside of the concrete. The appearance of cracks will accelerate further corrosion of the steel bars, forming a vicious cycle. This situation significantly reduces the durability and reliability of the substation concrete frame, shortens its service life, and poses a serious threat to the safe and stable operation of the substation. Therefore, real-time detection of the degree of steel corrosion in concrete structures has become an important aspect of evaluating the residual mechanical properties of concrete structures.

[0004] Since the steel bars are located inside the concrete and the substation is important and special, it is necessary to avoid damaging the integrity and durability of the concrete frame during the inspection process. Therefore, non-destructive testing technology has become the first choice. At present, non-destructive testing methods include electromagnetic method, macrocurrent method and ultrasonic testing method. Among them, ultrasonic testing has shown broad application prospects in the field of steel corrosion detection due to its high efficiency and precision. However, in practical applications, ultrasonic testing technology still faces certain technical difficulties and needs to be further optimized and improved. First, in areas with lighter corrosion thickness, the ultrasonic signal attenuation is small and the reflected signal intensity is high; while in areas with severe corrosion thickness, the ultrasonic signal attenuation is large and the reflected signal intensity is low; the corrosion thickness and degree of different corrosion sites vary greatly, and a single ultrasonic testing parameter is difficult to fully evaluate the overall corrosion condition of the component. Secondly, the geometric shape and spatial position of the corrosion site are complex and changeable, which has a significant impact on the propagation and reflection characteristics of the ultrasonic wave; the irregular shape of the corrosion site will cause the scattering and diffraction of the ultrasonic wave, affecting the accuracy of the detection signal; the change in the spatial position of the corrosion site will cause the change of the ultrasonic incident angle and propagation path, thereby affecting the reliability of the detection data. Finally, different types of corrosion products, such as iron oxide and iron hydroxide, have different acoustic properties. The distribution of corrosion products on the surface of components is also uneven. The types and distribution characteristics of corrosion products will affect the propagation and reflection behavior of ultrasonic waves on the surface of components, and thus affect the accuracy of corrosion assessment. In summary, in the current ultrasonic-based substation reinforced concrete corrosion thickness assessment method, a single ultrasonic detection parameter is difficult to fully assess the overall corrosion condition of the component. The geometric shape and spatial position of the rusted part are complex and changeable, and the types and distribution characteristics of corrosion products will affect the propagation and reflection behavior of ultrasonic waves on the surface of components. The combined influence leads to inaccurate assessment results. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a method and system for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound. The present invention determines the ultrasonic parameter combination by matching according to the geometric shape information and spatial position information of the corrosion area to be detected, thereby realizing real-time adjustment of the ultrasonic parameters when adapting to different corrosion positions; the type of corrosion in the corrosion area to be detected is obtained according to the spectral characteristic parameters extracted from the ultrasonic detection data and the preset support vector machine classification model; the corrosion thickness and the corrosion degree grade are obtained according to the corrosion type and the preset corrosion degree assessment model, thereby solving the influence of different types and distribution characteristics of corrosion products on the propagation and reflection of ultrasonic waves on the surface of components.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0007] In a first aspect, the present invention provides a method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound, comprising:

[0008] Acquire geometric shape information and spatial position information of the rusted area to be detected; determine the ultrasonic parameter combination by matching according to the geometric shape information and the spatial position information;

[0009] Based on the ultrasonic parameter combination, ultrasonic testing is performed on the rusted area to be tested to obtain ultrasonic testing data;

[0010] Extracting the frequency spectrum characteristic parameters of the ultrasonic detection data; obtaining the type of rust in the rust area to be detected according to the frequency spectrum characteristic parameters and a preset support vector machine classification model;

[0011] According to the type of rust and a preset rust degree assessment model, the rust thickness and rust degree grade are obtained; wherein the rust degree assessment model is a machine learning model.

[0012] Further, the three-dimensional model data of the rust target to be detected is obtained, and the geometric shape information of the rust area to be detected is obtained according to the three-dimensional model data; the ultrasonic detection data is registered with the three-dimensional model data to obtain the spatial position information of the rust target to be detected;

[0013] According to the geometric shape information and the spatial position information, in combination with a preset ultrasonic detection parameter library, a pattern matching algorithm is used to obtain the ultrasonic parameter combination; wherein the parameter combination includes ultrasonic frequency and probe type.

[0014] Furthermore, the ultrasonic detection data and the three-dimensional model data are point cloud registered, and the rigid body transformation matrix is ​​solved by the least square method to obtain the ultrasonic detection data in the three-dimensional space; the ultrasonic detection data in the three-dimensional space is image segmented to obtain the spatial position information of the rusted area to be detected in the three-dimensional space;

[0015] A regional growing algorithm is adopted, and the reflection characteristics of the ultrasonic retrieval data of the rust area are used as seed points to extract the contour information of the rust area and obtain the geometric shape of the rust area. According to the geometric shape of the rust area and the three-dimensional model data, the spatial position information of the rust area to be detected in the three-dimensional space is obtained through the triangulation principle.

[0016] Furthermore, the ultrasonic detection data is decomposed into several levels by wavelet transform, and the decomposed ultrasonic detection data is selected and subjected to soft threshold denoising; the denoised data is normalized, and the normalized data is subjected to spectrum analysis by Fourier transform to extract spectrum feature parameters.

[0017] Furthermore, the Euclidean distance measurement method is used to calculate the similarity between the spectral feature parameters and the spectral features of each rust product in a preset spectral feature database; a preset number of rust products whose similarities are ranked in the front are selected as candidates; the spectral feature parameters of the candidate rust products and the spectral feature parameters of the ultrasonic detection data are input into the support vector machine classification model to obtain the rust type of the rust area to be detected.

[0018] Furthermore, there are multiple corrosion degree assessment models, each of which corresponds to a corrosion product. If a certain type of corrosion product does not have a corresponding corrosion degree assessment model, the Euclidean distance measurement method is used to calculate the similarity between this type of corrosion product and the spectral characteristics of the corrosion product corresponding to the constructed corrosion degree assessment model, and the corrosion degree assessment model corresponding to the corrosion product with the highest similarity is selected for evaluation.

[0019] In a second aspect, the present invention also provides a substation reinforced concrete corrosion thickness assessment system based on ultrasound, comprising:

[0020] The ultrasonic parameter combination determination module is configured to: obtain geometric shape information and spatial position information of the rust area to be detected; determine the ultrasonic parameter combination by matching according to the geometric shape information and the spatial position information;

[0021] The ultrasonic detection module is configured to: perform ultrasonic detection on the rusted area to be detected based on the ultrasonic parameter combination to obtain ultrasonic detection data;

[0022] The classification module is configured to: extract the frequency spectrum characteristic parameters of the ultrasonic detection data; and obtain the type of rust in the rust area to be detected according to the frequency spectrum characteristic parameters and a preset support vector machine classification model;

[0023] The evaluation module is configured to obtain the rust thickness and rust degree grade according to the rust type and a preset rust degree evaluation model; wherein the rust degree evaluation model is a machine learning model.

[0024] In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the ultrasonic-based substation reinforced concrete corrosion thickness assessment method described in the first aspect.

[0025] In a fourth aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein when the processor executes the program, the steps of the ultrasonic-based substation reinforced concrete corrosion thickness assessment method described in the first aspect are implemented.

[0026] In a fifth aspect, the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the ultrasonic-based substation reinforced concrete corrosion thickness assessment method described in the first aspect are implemented.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] In the present invention, according to the geometric shape information and spatial position information of the rusted area to be detected, the ultrasonic parameter combination is determined by matching, so that the real-time adjustment of the ultrasonic parameters when adapting to different rusted parts is realized; according to the spectral characteristic parameters of the extracted ultrasonic detection data and the preset support vector machine classification model, the type of rust in the rusted area to be detected is obtained; according to the type of rust and the preset rust degree assessment model, the rust thickness and the rust degree grade are obtained, and the influence of different types and distribution characteristics of rust products on the propagation and reflection of ultrasonic waves on the surface of components is solved; by adjusting the ultrasonic detection parameters to fully reflect the overall rust condition, considering the geometric shape and spatial position of the rusted part, and comprehensively improving the assessment according to the type of rust, the assessment method can adapt to complex rusted parts, improve the quality of detection data, realize the comprehensive and accurate assessment of the rust degree of the reinforced concrete structure of the substation, and provide a reliable basis for the maintenance and safety assessment of the reinforced concrete structure of the substation. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0030] Figure 1 This is a flow chart of the method of Example 1 of the present invention. DETAILED DESCRIPTION

[0031] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0032] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0033] Embodiment 1:

[0034] Ultrasonic testing is a non-destructive testing method that uses ultrasonic technology to perform testing. Non-destructive testing is a means of testing the surface and internal quality of the inspected component without damaging the workpiece or the working state of the raw material. The transmission loss of ultrasonic waves in solids is very small and the detection depth is large. Ultrasonic waves will reflect and refract at heterogeneous interfaces, especially they cannot pass through gas-solid interfaces. If there are defects or inclusions such as pores, cracks, and stratification in the metal, the ultrasonic wave will be fully or partially reflected when it propagates to the interface between the metal and the defect. The reflected ultrasonic wave is received by the probe and processed by the circuit inside the instrument, and waveforms of different heights and certain spacing will be displayed on the instrument's fluorescent screen.

[0035] As described in the background technology, in practical applications, ultrasonic detection technology faces some technical problems that need to be solved urgently. First, the degree of corrosion of different rusted parts varies greatly, and a single ultrasonic detection parameter is difficult to comprehensively evaluate the overall corrosion condition of the component. In parts with a lighter degree of corrosion, the ultrasonic signal attenuation is small and the reflected signal intensity is high; while in parts with a serious degree of corrosion, the ultrasonic signal attenuation is large and the reflected signal intensity is low; how to establish a reasonable corrosion degree assessment model based on the ultrasonic detection data of different parts to achieve an accurate assessment of the overall corrosion condition of the component is a problem that needs to be solved urgently. Secondly, the geometric shape and spatial position of the rusted parts are complex and changeable, which has a significant impact on the propagation and reflection characteristics of ultrasonic waves. The irregular shape of the rusted parts will cause scattering and diffraction of ultrasonic waves, affecting the accuracy of the detection signal; the change in the spatial position of the rusted parts will cause changes in the ultrasonic incident angle and propagation path, thereby affecting the reliability of the detection data; how to optimize the ultrasonic detection parameters for rusted parts of different shapes and positions to improve the accuracy and reliability of the detection data is another problem that needs to be solved urgently. Finally, the impact of the type and distribution of rust products on ultrasonic testing results cannot be ignored; different types of rust products, such as iron oxide and iron hydroxide, have quite different acoustic properties; the distribution of rust products on the surface of components is also uneven; the type and distribution characteristics of rust products will affect the propagation and reflection behavior of ultrasonic waves on the surface of components, and thus affect the accuracy of rust degree assessment; how to identify the type of rust products through ultrasonic testing data and consider the impact of their distribution characteristics on the test results is the key to improving the accuracy of rust degree assessment.

[0036] It is difficult to comprehensively evaluate the overall corrosion condition of components based on a single ultrasonic testing parameter. The geometric shape and spatial position of the corrosion site are complex and changeable, and the type and distribution characteristics of the corrosion products will affect the propagation and reflection behavior of the ultrasonic wave on the surface of the component. This combined influence leads to inaccurate evaluation results, such as Figure 1As shown, this embodiment provides a method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound, which is suitable for complex corrosion sites, improves the quality of detection data, realizes a comprehensive and accurate assessment of the degree of corrosion of metal components, and provides a reliable basis for component maintenance and safety assessment; the method can be used to assess the corrosion thickness of reinforced concrete structures in substations, and the method includes:

[0037] S1. Obtain geometric shape information and spatial position information of the rusted area to be detected, and obtain an ultrasonic parameter combination according to the geometric shape information and the spatial position information.

[0038] S2. Perform ultrasonic testing on the rusted area to be tested based on the ultrasonic parameter combination to obtain ultrasonic testing data.

[0039] S3. Input the frequency spectrum characteristic parameters of the ultrasonic detection data into a support vector machine classification model, and output the judgment result of the rust type in the rust area to be detected; wherein the support vector machine classification model is obtained based on the first training set training, and the first training set includes known rust types and corresponding ultrasonic frequency spectrum characteristic parameters.

[0040] S4. According to the result of the corrosion type judgment, the ultrasonic detection data is input into the corresponding corrosion degree assessment model to obtain the corrosion thickness and corrosion degree grade; wherein the corrosion degree assessment model is constructed based on a machine learning model and is obtained through training based on a second training set, and the second training set includes the corrosion historical ultrasonic detection data corresponding to the corrosion type judgment result, the historical degree assessment results and the corrosion thickness of the historical detection.

[0041] In step S1, obtaining an ultrasonic parameter combination according to geometric shape information and spatial position information includes:

[0042] S1.1. According to the 3D model data of the rust target to be detected, the geometric shape information of the rust area to be detected is obtained; optionally, a laser scanner is used to capture the 3D model data of the surface of the component in the rust area to be detected. The rust target to be detected refers to a metal component, and the rust area to be detected can be the area to be inspected on the metal component of the reinforced concrete structure of the substation.

[0043] S1.2. Align the ultrasonic detection data with the three-dimensional model data of the rusted area to be detected to obtain the spatial position information of the rusted target to be detected.

[0044] S1.2.1. Use the ICP (Iterative Closest Point) algorithm to perform point cloud registration between the ultrasonic detection data and the three-dimensional model data, and optimize the rigid body transformation matrix through the least squares method to obtain the ultrasonic detection data in the three-dimensional space.

[0045] S1.2.2. Perform image segmentation processing on the ultrasonic detection data in the three-dimensional space to obtain the spatial position information of the rusted area to be detected in the three-dimensional space.

[0046] S1.2.2.1. Use the regional growing algorithm, take the reflection characteristics of the ultrasonic retrieval data of the rusted area as the seed point, extract the contour information of the rusted area, and obtain the geometric shape of the rusted area.

[0047] Optionally, a region growing algorithm is used to extract contour information of the rusted area using the reflection features of the ultrasonic retrieval data of the rusted area as seed points, specifically including:

[0048] 1) Determine seed points: Based on the reflection characteristics of the ultrasonic retrieval data, identify the characteristic points of the rusted area as the initial seed points. These seed points are usually areas with abnormal reflection intensity, indicating the possible presence of rust.

[0049] 2) Define the growth criterion: Set a growth criterion based on the similarity between the reflection characteristics of the pixel points in the ultrasonic data and the seed points. A threshold is preset, and only when the reflection characteristics of the adjacent pixels and the characteristics of the seed points are within the threshold range, the pixel will be added to the growth area.

[0050] 3) Initialize the growing region: Create a marker image of the same size as the ultrasound image to record which pixels have been added to the growing region. Initially, except for the seed point, all other pixels are not marked.

[0051] 4) Growth process: Starting from the seed point, check the pixels in its neighborhood, using an 8-neighborhood or a 4-neighborhood. For each neighborhood pixel, calculate the similarity between its reflection feature and the seed point.

[0052] If the reflectance feature of the neighborhood pixel is within the predefined growth criterion threshold, it is added to the growth region and marked at the corresponding position in the marker image.

[0053] 5) Iterative growth: Take the newly added pixel points in the growth area as new seed points, repeat step 4), and continue the growth process until no more pixels meet the growth criteria or the preset termination condition is reached.

[0054] 6) Termination condition: A termination condition is preset, which can be based on the size, shape or density of the growth area. Once the termination condition is met, the growth process stops.

[0055] 7) Extracting contour information: After the growth is completed, the marked image is post-processed to extract the contour information of the growth area.

[0056] S1.2.2.2. Based on the geometric shape of the rusted area and the three-dimensional model data, the spatial position information of the rusted area to be detected in three-dimensional space is obtained through the principle of triangulation.

[0057] Optionally, first, the ultrasonic detection data is registered with the pre-built three-dimensional model data of the target area, and the ICP algorithm is used for point cloud matching. The rigid body transformation matrix is ​​optimized and solved by the least squares method to achieve accurate alignment between the ultrasonic detection data and the three-dimensional model, and obtain ultrasonic detection data in three-dimensional space. Then, the registered ultrasonic detection data is processed by image segmentation to extract the contour information of the rusted area. According to the geometric shape and three-dimensional model data of the rusted area, the precise position coordinates of the rusted area in three-dimensional space are calculated by the triangulation principle, and the positioning accuracy can reach 1mm. Finally, the geometric shape and position coordinate information of the rusted area are visualized, and the Marching Cubes algorithm is used to reconstruct the rusted area in three dimensions and highlight it in red to generate an intuitive rust distribution map.

[0058] S1.3. According to the geometric shape information and the spatial position information, and in combination with the preset ultrasonic detection parameter library, a pattern matching algorithm is used to obtain an ultrasonic parameter combination; wherein the parameter combination includes an ultrasonic frequency and a probe type.

[0059] Optionally, the pattern matching algorithm is an operation of string processing in a data structure, mainly including the naive pattern matching algorithm, the KMP (Knuth-Morris-Pratt) algorithm, the BM (Bad Match) algorithm and the Rabin-Karp algorithm.

[0060] This embodiment uses a naive pattern matching algorithm to select the most suitable ultrasonic frequency and probe type from a preset ultrasonic detection parameter library. The specific steps are as follows:

[0061] 1) Build a parameter library: Build a parameter library containing different ultrasonic frequencies and probe types. The library contains the usage characteristic parameters and acoustic performance parameters of each probe, such as center frequency, pulse width, relative bandwidth, near field length, beam diameter and diffusion angle.

[0062] 2) Define matching rules: Define matching rules based on the geometric shape information and spatial position information of the detected object. These rules are used as pattern strings to match the entries in the parameter library.

[0063] 3) Pattern string construction: Construct a pattern string that contains the characteristics of the required ultrasonic frequency and probe type.

[0064] 4) Traverse the parameter library: traverse each entry in the parameter library, regard each entry as a text string, and match the pattern string with each entry.

[0065] 4) Perform naive pattern matching: For each entry in the parameter library, use the naive pattern matching algorithm to match. Starting from the first entry in the parameter library, compare the pattern string with the entry character by character.

[0066] 5) Matching process: Use two nested loops to traverse the pattern string and parameter library entries. The outer loop traverses each entry in the parameter library, and the inner loop compares the features of the pattern string and the entry one by one. If an unmatched feature is found, the inner loop will immediately exit and continue to compare the next entry. If the entire pattern string matches an entry, the entry is recorded as a potential match result.

[0067] 6) Result screening: According to the degree of matching, filter out the entries that best meet the matching rules.

[0068] In step S2, after acquiring the ultrasonic detection data, the collected ultrasonic data is subjected to wavelet denoising and normalization processing to improve the signal-to-noise ratio of the signal. Based on wavelet packet decomposition, frequency domain characteristic parameters reflecting the degree of corrosion are extracted, such as wavelet packet energy spectrum entropy and wavelet packet singular values.

[0069] S2.1. Use wavelet transform to decompose the ultrasonic detection data into several levels, select the decomposed ultrasonic detection data and perform soft threshold denoising.

[0070] S2.2. The denoised data is normalized, and Fourier transform is used to perform spectrum analysis on the normalized data to extract spectrum characteristic parameters; wherein the spectrum characteristic parameters include the spectrum center of gravity, the spectrum peak, and the frequency corresponding to the spectrum peak.

[0071] Optionally, the signal is decomposed into 5 levels by wavelet transform method, and the high frequency coefficients of levels 3 to 5 are selected and soft threshold denoising is performed. The denoised signal is then normalized and the signal amplitude is mapped to the interval [0, 1]. Next, the 512-point fast Fourier transform is used to perform spectrum analysis on the preprocessed signal to extract the frequency domain characteristic parameters of the signal, including the spectrum center of gravity, spectrum peak and the frequency corresponding to the spectrum peak.

[0072] In step S3, the frequency spectrum characteristic parameters of the ultrasonic detection data are input into the support vector machine classification model, and the output of the rust type judgment result in the rust area to be detected includes:

[0073] S3.1. Pre-establish a spectral feature database of corrosion products, use the Euclidean distance measurement method to calculate the similarity between the spectral feature parameters and the spectral features of each corrosion product in the database, and select the first several corrosion products with the highest similarity as candidates.

[0074] Optionally, the Euclidean distance measurement method is used to calculate the similarity between the spectrum feature parameters and the spectrum features of each corrosion product in the database, including:

[0075] Extract characteristic parameters from the spectrum data of the corrosion product, construct the extracted characteristic parameters into a characteristic vector, and use the Euclidean distance formula to calculate the distance between the characteristic vector of the spectrum characteristic parameters and the characteristic vector of each sample in the database. For two n-dimensional vectors a and b, the calculation formula of the Euclidean distance d is:

[0076]

[0077] Among them, x 1k and x 2k are the values ​​of vectors a and b in the kth dimension respectively.

[0078] S3.2. Input the candidate spectral characteristic parameters of the rust product and the spectral characteristic parameters of the ultrasonic detection data into the support vector machine classification model to obtain the judgment result of the rust type in the rust area to be detected.

[0079] Optionally, based on a pre-established spectral feature database containing 100 common rust products, the Euclidean distance measurement method is used to calculate the similarity between the spectral feature parameters of the current signal and the spectral features of each rust product in the database, and the top five rust products with the highest similarity are selected as candidates. Finally, the spectral feature parameters of the candidate rust products and the spectral feature parameters of the ultrasonic detection data are input into a multi-classification model based on a support vector machine (SVM) for discrimination. The model uses an RBF (Radial Basis Function) kernel function, and the classification accuracy after cross-validation optimization can reach more than 95%.

[0080] In step S4, optionally, the machine learning model is a support vector machine and / or a random forest. The support vector machine algorithm is used to construct a corrosion degree assessment model, using a radial basis kernel function, a penalty factor C=10, a kernel function parameter g=1, and the optimal parameters are determined by 5-fold cross validation. The machine learning model is trained using the training set data, and the prediction accuracy on the test set can reach more than 90%. There are several types of corrosion degree assessment models, each of which corresponds to a type of corrosion product. If a certain type of corrosion product does not have a corresponding corrosion degree assessment model, the Euclidean distance measurement method is used to calculate the similarity of the spectral characteristics of this type of corrosion product and the corrosion product corresponding to the constructed corrosion degree assessment model, and the corrosion degree assessment model corresponding to the corrosion product with the highest similarity is selected. In view of the complex and changeable geometric shape and spatial position of the corrosion site, this embodiment adopts adaptive ultrasonic detection technology to adapt to different corrosion sites by adjusting ultrasonic parameters in real time; the type of corrosion product is identified by ultrasonic spectrum analysis, and the detection data obtained by different parameters is comprehensively utilized by a multi-scale fusion strategy. Finally, a corrosion degree assessment model is established using a machine learning algorithm to accurately assess the overall corrosion status of the component. This embodiment can adapt to complex corrosion sites, improve the quality of detection data, and achieve a comprehensive and accurate assessment of the degree of corrosion of metal components, providing a reliable basis for component maintenance and safety assessment.

[0081] Embodiment 2:

[0082] This embodiment provides a substation reinforced concrete corrosion thickness assessment system based on ultrasound, including:

[0083] The ultrasonic parameter combination determination module is configured to: obtain geometric shape information and spatial position information of the rust area to be detected; determine the ultrasonic parameter combination by matching according to the geometric shape information and the spatial position information;

[0084] The ultrasonic detection module is configured to: perform ultrasonic detection on the rusted area to be detected based on the ultrasonic parameter combination to obtain ultrasonic detection data;

[0085] The classification module is configured to: extract the frequency spectrum characteristic parameters of the ultrasonic detection data; and obtain the type of rust in the rust area to be detected according to the frequency spectrum characteristic parameters and a preset support vector machine classification model;

[0086] The evaluation module is configured to obtain the rust thickness and rust degree grade according to the rust type and a preset rust degree evaluation model; wherein the rust degree evaluation model is a machine learning model.

[0087] The working method of the system is the same as the ultrasonic-based method for assessing the corrosion thickness of reinforced concrete in a substation in Example 1, and will not be described in detail here.

[0088] Embodiment 3:

[0089] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound described in Example 1 are implemented.

[0090] Embodiment 4:

[0091] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the program, the steps of the ultrasonic-based substation reinforced concrete corrosion thickness assessment method described in Example 1 are implemented.

[0092] Embodiment 5:

[0093] This embodiment provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound described in Example 1 are implemented.

[0094] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound, characterized in that: include: Acquire geometric shape information and spatial position information of the rusted area to be detected; determine the ultrasonic parameter combination by matching according to the geometric shape information and the spatial position information; Based on the ultrasonic parameter combination, ultrasonic testing is performed on the rusted area to be tested to obtain ultrasonic testing data; Extracting the frequency spectrum characteristic parameters of the ultrasonic detection data; obtaining the type of rust in the rust area to be detected according to the frequency spectrum characteristic parameters and a preset support vector machine classification model; According to the type of rust and a preset rust degree assessment model, the rust thickness and rust degree grade are obtained; wherein the rust degree assessment model is a machine learning model.

2. The method for evaluating the corrosion thickness of reinforced concrete in a substation based on ultrasound according to claim 1, characterized in that: Acquire the three-dimensional model data of the rust target to be detected, and obtain the geometric shape information of the rust area to be detected according to the three-dimensional model data; align the ultrasonic detection data with the three-dimensional model data to obtain the spatial position information of the rust target to be detected; According to the geometric shape information and the spatial position information, in combination with a preset ultrasonic detection parameter library, a pattern matching algorithm is used to obtain the ultrasonic parameter combination; wherein the parameter combination includes ultrasonic frequency and probe type.

3. The method for evaluating the corrosion thickness of reinforced concrete in a substation based on ultrasound according to claim 2, characterized in that: Performing point cloud registration on the ultrasonic detection data and the three-dimensional model data, and solving the rigid body transformation matrix by the least square method to obtain the ultrasonic detection data in the three-dimensional space; performing image segmentation processing on the ultrasonic detection data in the three-dimensional space to obtain the spatial position information of the rusted area to be detected in the three-dimensional space; A regional growing algorithm is adopted, and the reflection characteristics of the ultrasonic retrieval data of the rust area are used as seed points to extract the contour information of the rust area and obtain the geometric shape of the rust area. According to the geometric shape of the rust area and the three-dimensional model data, the spatial position information of the rust area to be detected in the three-dimensional space is obtained through the triangulation principle.

4. The method for evaluating the corrosion thickness of reinforced concrete in a substation based on ultrasound according to claim 1, characterized in that: The ultrasonic detection data is decomposed into several levels by wavelet transform, and the decomposed ultrasonic detection data is selected and subjected to soft threshold denoising; the denoised data is normalized, and the normalized data is subjected to spectrum analysis by Fourier transform to extract spectrum feature parameters.

5. The method for evaluating the corrosion thickness of reinforced concrete in a substation based on ultrasound according to claim 1, characterized in that: Using the Euclidean distance measurement method, the similarity between the spectrum feature parameters and the spectrum features of each corrosion product in the preset spectrum feature database is calculated; A preset number of multiple rust products whose similarity ranks in front are selected as candidates; the spectral characteristic parameters of the candidate rust products and the spectral characteristic parameters of the ultrasonic detection data are input into the support vector machine classification model to obtain the rust type of the rust area to be detected.

6. The method for evaluating the corrosion thickness of reinforced concrete in a substation based on ultrasound according to claim 5, characterized in that: There are many types of corrosion degree assessment models, and each corrosion degree assessment model corresponds to a corrosion product. If a certain type of corrosion product does not have a corresponding corrosion degree assessment model, the Euclidean distance measurement method is used to calculate the similarity between this type of corrosion product and the spectral characteristics of the corrosion product corresponding to the constructed corrosion degree assessment model, and the corrosion degree assessment model corresponding to the corrosion product with the highest similarity is selected for evaluation.

7. An ultrasonic-based substation reinforced concrete corrosion thickness assessment system, characterized in that: include: The ultrasonic parameter combination determination module is configured to: obtain geometric shape information and spatial position information of the rust area to be detected; determine the ultrasonic parameter combination by matching according to the geometric shape information and the spatial position information; The ultrasonic detection module is configured to: perform ultrasonic detection on the rusted area to be detected based on the ultrasonic parameter combination to obtain ultrasonic detection data; The classification module is configured to: extract the frequency spectrum characteristic parameters of the ultrasonic detection data; and obtain the type of rust in the rust area to be detected according to the frequency spectrum characteristic parameters and a preset support vector machine classification model; The evaluation module is configured to obtain the rust thickness and rust degree grade according to the rust type and a preset rust degree evaluation model; wherein the rust degree evaluation model is a machine learning model.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound as described in any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the program, the steps of the method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound are implemented as described in any one of claims 1 to 6.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the method for assessing the corrosion thickness of reinforced concrete in a substation based on ultrasound are implemented as described in any one of claims 1 to 6.

Images