A method and device for detecting concrete corrosion state

Abstract

The present application relates to the field of chlorine detection in concrete and discloses a method and apparatus for detecting the corrosion state of concrete. The method comprises: obtaining response data for a first time period and a second time period of a concrete sample to be tested; determining concentration response data based on the response data for the first time period; and determining depth response data based on the response data for the second time period; in response to a concentration detection instruction, generating the chlorine concentration of the concrete sample to be tested based on the concentration response data and the depth response data under a first preset weight; in response to a depth detection instruction, generating the corrosion depth of the concrete sample to be tested based on the concentration response data and the depth response data under a second preset weight; and generating detection information for characterizing the corrosion state of the concrete sample to be tested based on the chlorine concentration and the corrosion depth. The technical solution provided by the present application can achieve accurate prediction of the chlorine corrosion state in concrete.

Description

Technical Field

[0001] The present application relates to the technical field of concrete chlorine element detection, and in particular to a method and equipment for detecting the corrosion state of concrete. Background Art

[0002] Chloride attack is an important factor affecting the durability of concrete structures. Currently, PGNAA technology is commonly used to detect the chlorine content in concrete structures. However, since the PGNAA detection system is relatively fixed, chlorine analysis is usually performed on the PGNAA system after concrete sampling, which has a certain degree of randomness.

[0003] To more accurately determine the chlorine corrosion status of concrete structures, the application of PGNAA (Physical Navigation Analysis) to concrete element analysis on mobile devices has been proposed. However, the detection of corrosion information describing the chlorine corrosion state in concrete currently relies solely on chlorine content detection. Relying solely on chlorine content as corrosion information to determine the corrosion state can lead to certain errors. Therefore, accurately predicting the chlorine corrosion state in concrete has become a key research focus in the field of concrete chlorine detection. Summary of the Invention

[0004] The present application provides a method and device for detecting the corrosion state of concrete, which can accurately predict the corrosion state of chlorine in concrete.

[0005] In a first aspect, the present application provides a method for detecting a concrete corrosion state. The method includes: acquiring response data of a first time period and response data of a concrete sample to be tested, determining concentration response data based on the response data of the first time period, and determining depth response data based on the response data of the second time period; in response to a concentration detection instruction, generating a chlorine concentration of the concrete sample to be tested based on the concentration response data and the depth response data under a first preset weight; in response to a depth detection instruction, generating a corrosion depth of the concrete sample to be tested based on the concentration response data and the depth response data under a second preset weight; and generating detection information based on the chlorine concentration and the corrosion depth, wherein the detection information is used to characterize the corrosion state of the concrete sample to be tested.

[0006] In one possible embodiment, obtaining the response data of the concrete sample to be tested for the first time period and the response data of the concrete sample to be tested for the second time period includes: using a detector to approach the concrete sample to be tested at a uniform speed, and obtaining multiple detector response values ​​received by the detector in different time periods during the uniform approach process; using the multiple detector response values ​​of the detector in the first time period as the response data for the first time period, and using the multiple detector response values ​​of the detector in the second time period as the response data for the second time period.

[0007] In one possible embodiment, determining concentration response data based on the response data of the first time period, and determining depth response data based on the response data of the second time period includes: obtaining the response integral of the response data of the first time period in the first time period, using the response integral as the concentration response data, obtaining the response change rate of the response data of the second time period in the second time period, and using the response change rate as the depth response data, wherein the response time interval represented by the second time period is earlier than the response time interval represented by the first time period.

[0008] In one possible implementation, the chlorine concentration of the concrete sample to be tested is generated by a concentration detection model. The concentration detection model generates the chlorine concentration of the concrete sample to be tested in the following manner: the concentration detection model generates a concentration mapping parameter based on the first preset weight and the concentration response data and the depth response data, where the first preset weight is used to represent the concentration detection model's attention to the concentration response data; the concentration mapping parameter is mapped to a high-dimensional concentration space to obtain a corresponding high-dimensional concentration mapping vector; and the chlorine concentration of the concrete sample to be tested is generated based on the high-dimensional concentration mapping vector and a concentration correlation function.

[0009] In one possible implementation, the corrosion depth of the concrete sample to be tested is generated by a depth detection model. The depth detection model generates the corrosion depth of the concrete sample to be tested in the following manner: the depth detection model generates a depth mapping parameter based on the concentration response data and the depth response data based on the second preset weight, where the second preset weight is used to represent the attention of the depth detection model to the depth response data; the depth mapping parameter is mapped to a deep high-dimensional space to obtain a corresponding high-dimensional depth mapping vector; and the corrosion depth of the concrete sample to be tested is generated based on the high-dimensional depth mapping vector and a depth correlation function.

[0010] In one possible embodiment, the concentration detection model is a pre-trained SVR model; the concentration detection model is trained in the following manner: concentration response data and depth response data of multiple groups of concrete training samples are obtained, and a preset concentration and preset depth of the chlorine element in each group of concrete simulation samples are obtained, wherein the preset concentration and preset depth of the chlorine element in each group of concrete simulation samples are different; based on the concentration response data and the depth response data of each group of concrete training samples, a concentration mapping parameter is generated using the first preset weight; based on the concentration mapping parameters and the preset concentration of each group of concrete training samples, a concentration correlation function of the concentration detection model is generated, and the concentration detection model is trained using a first SVR loss function.

[0011] In one possible embodiment, the depth detection model is a pre-trained SVR model; the depth detection model is trained in the following manner: concentration response data and depth response data of multiple groups of concrete training samples are obtained, and a preset concentration and preset depth of the chlorine element in each group of concrete simulation samples are obtained, wherein the preset concentration and preset depth of the chlorine element in each group of concrete simulation samples are different; based on the concentration response data and the depth response data of each group of concrete training samples, depth mapping parameters are generated using the second preset weights; based on the depth mapping parameters and the preset depth of each group of concrete training samples, a depth correlation function of the depth detection model is generated, and the depth detection model is trained using a second SVR loss function.

[0012] In one possible embodiment, after constructing the concentration detection model, the method further includes: obtaining multiple groups of concrete prediction samples, obtaining the true concentration of chlorine in any concrete prediction sample, and generating a predicted concentration of chlorine in the concrete prediction sample based on the concentration detection model; generating a prediction error and an evaluation index of the concentration detection model based on the true concentration and the predicted concentration, and if the evaluation index does not meet a preset condition, repeatedly adjusting the first preset weight and the first SVR loss function according to the prediction error until the evaluation index of the concentration detection model meets the preset condition.

[0013] In one possible embodiment, after constructing the depth detection model, the method further includes: obtaining multiple groups of concrete prediction samples, obtaining the true depth of chlorine in any concrete prediction sample, and generating a predicted depth of chlorine in the concrete prediction sample based on the concentration detection model; generating a prediction error and an evaluation index of the depth detection model based on the true depth and the predicted depth, and if the evaluation index does not meet a preset condition, repeatedly adjusting the second preset weight and the second SVR loss function according to the prediction error until the evaluation index of the depth detection model meets the preset condition.

[0014] According to a second aspect of the present application, a device for detecting a concrete corrosion state is provided. The device includes: a data acquisition unit for acquiring first-period response data and second-period response data of a concrete sample to be tested, determining concentration response data based on the first-period response data, and determining depth response data based on the second-period response data; a data processing unit for inputting the concentration response data and the depth response data into a concentration detection model in response to a concentration detection instruction, wherein the concentration detection model outputs a chlorine concentration of the concrete sample to be tested based on a first preset weight; and a data processing unit for inputting the concentration response data and the depth response data into a depth detection model in response to a depth detection instruction, wherein the depth detection model outputs a corrosion depth of the concrete sample to be tested based on a second preset weight; and an information generation unit for generating detection information based on the chlorine concentration and the corrosion depth, wherein the detection information is used to characterize the corrosion state of the concrete sample to be tested.

[0015] A third aspect of the present application provides a computer device, comprising a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions. The computer device is used to implement the concrete corrosion state detection method described in the first aspect.

[0016] The technical solution provided by one or more embodiments of the present application dynamically analyzes the response data of different time periods through two pre-trained models, and can obtain more accurate corrosion information describing the corrosion state of concrete by chlorine. Specifically, the response data of the concrete sample to be tested at different time periods is dynamically obtained by a movable detector, and the first time period response data that can reflect the distribution of chlorine in the concrete and the second time period response data that can reflect the depth information of concrete corrosion are selected. The first time period response data and the second time period response data are subjected to targeted model detection to obtain the chlorine concentration and corrosion depth of the concrete sample to be tested. The chlorine concentration and corrosion depth are used as corrosion information to describe the corrosion state of chlorine in concrete, rather than just the chlorine content as corrosion information, and the accuracy of chlorine detection can be improved by dynamically adjusting the model weight according to the prediction error. Compared with the prior art, the depth prediction of the concrete corrosion state can be achieved by dynamically acquiring and analyzing the response information to obtain more accurate corrosion information.

[0017] It can be seen that the technical solution provided by this application can obtain deeper corrosion information and accurately predict the corrosion state of chlorine in concrete. At the same time, the dynamic detection of the detector can also improve the convenience of chlorine corrosion state detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A schematic diagram of a detection scenario for a vehicle-mounted PGNAA detection system provided in this application;

[0020] Figure 2 A schematic diagram of the steps of a method for detecting concrete corrosion status provided in an embodiment of the present application;

[0021] Figure 3 A schematic diagram of the three-dimensional relationship between chlorine concentration, corrosion depth, and response data provided in one embodiment of the present application;

[0022] FIG4 (a) and FIG4 (b) are schematic diagrams of a timing sequence of response data provided by an embodiment of the present application;

[0023] FIG5( a ) and FIG5 ( b ) are exemplary diagrams of a detection scenario for detecting concrete corrosion status according to an embodiment of the present application;

[0024] FIG6 (a) and FIG6 (b) are schematic diagrams showing the mapping relationship between response data and corrosion characteristics provided in one embodiment of the present application;

[0025] FIG7 (a) and FIG7 (b) are schematic diagrams of prediction errors of chlorine concentration and corrosion depth provided in one embodiment of the present application;

[0026] Figure 8 A schematic structural diagram of a concrete corrosion state detection device provided in one embodiment of the present application;

[0027] Figure 9 A schematic structural diagram of a computer device provided in accordance with one embodiment of the present application.

[0028] Description of Reference Numerals

[0029] 10-vehicle bracket, 11-chlorine element detector, 12-rotating device, 13-neutron source, 14-concrete sample, 15-corroded concrete, 21-detector, 22-corroded area, 23-concrete sample, 24-reaction area. DETAILED DESCRIPTION

[0030] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of this application.

[0031] In addition, the descriptions of "first", "second", etc. in this application are for descriptive purposes only and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the embodiments of the present application, unless otherwise specified, "multiple" means two or more. In addition, the use of "based on" or "according to" means openness and inclusiveness, because the process, steps, calculations or other actions "based on" or "according to" one or more of the conditions or values ​​can be based on additional conditions or values ​​beyond the described values ​​in practice.

[0032] PGNAA technology is an effective non-destructive elemental analysis method. Due to its high sensitivity to chlorine in concrete, PGNAA is often used to detect the chlorine content in concrete. Specifically, the PGNAA detection system emits neutrons to irradiate the sample. The neutrons are captured by the atomic nuclei in the sample and de-excited, emitting gamma rays. The energy and intensity of these gamma rays are related to the type and number of atomic nuclei. The chlorine detector configured in the PGNAA detection system captures the gamma rays and performs characteristic analysis to obtain the chlorine detector's response data. Quantitative analysis of the characteristic peaks and net counts of the response data can determine the composition and content of the chlorine in the sample, thus enabling the detection of the chlorine content.

[0033] Generally speaking, due to the relatively fixed nature of the PGNAA detection system, the aforementioned PGNAA detection system is used to test a small number of concrete samples after sampling. Since the chlorine detector can only determine the composition and content of chlorine within the sample, this detection method can only determine the chlorine content of the concrete sample, which may be subject to certain occasional errors. Furthermore, evaluating the corrosion state of concrete based solely on chlorine content is not accurate and comprehensive, and may be subject to certain judgment errors. Therefore, a method has been proposed for dynamic testing of concrete samples using a mobile PGNAA detection system. This method can obtain chlorine response data at different locations, thereby obtaining deeper chlorine information about the concrete sample.

[0034] See also Figure 1 The embodiments of the present application provide a scenario example of using a movable PGNAA detection system to perform dynamic detection on concrete samples.

[0035] Figure 1 The figure shows a detection schematic diagram of a vehicle-mounted PGNAA detection system. The PGNAA detection system is connected to a mobile device via a vehicle-mounted bracket 10. The PGNAA detection system is equipped with four chlorine detectors 11 and a neutron source 13. The four chlorine detectors 11 are evenly distributed at the bottom of the PGNAA detection system and are respectively connected to four rotating devices 12 on the PGNAA detection system. When detecting chlorine in concrete, neutrons emitted by the neutron source 13 react with corroded concrete 15 to generate gamma rays, which are captured by the chlorine detectors for characteristic analysis. During this process, the chlorine detector 11, connected to the vehicle-mounted bracket and the rotating mechanism, can adjust the distance and angle between it and the concrete sample 14, thereby obtaining chlorine response data that changes with distance. By uniformly approaching the concrete sample 14, the changes in the response data of the concrete sample 14 over different time periods can be recorded, achieving dynamic detection of the concrete sample 14, thereby providing a data basis for detecting chlorine concentration, corrosion depth, or other deep-level corrosion information.

[0036] The above description is merely an example scenario provided in the specification and is not intended to limit the present invention. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0037] In view of this, based on the above-mentioned movable PGNAA detection system, the embodiment of the present application provides a concrete corrosion status detection method and equipment, which can obtain more accurate chlorine element concentration and corrosion depth as corrosion information by analyzing the response data of different time periods, and determine the corrosion status of concrete through corrosion information at a deeper level than the chlorine element content, thereby realizing accurate prediction of the chlorine element corrosion status in concrete.

[0038] See also Figure 2 In one embodiment of the present application, a method for detecting the corrosion state of concrete is provided. The method may include the following steps:

[0039] S1: Acquire response data of a first time period and response data of a second time period of a concrete sample to be tested, determine concentration response data according to the response data of the first time period, and determine depth response data according to the response data of the second time period.

[0040] S3: In response to the concentration detection instruction, under a first preset weight, generate the chlorine concentration of the concrete sample to be tested according to the concentration response data and the depth response data.

[0041] S5: In response to the depth detection instruction, under the condition of a second preset weight, generating the corrosion depth of the concrete sample to be tested according to the concentration response data and the depth response data.

[0042] S7: Generate detection information based on the chlorine concentration and the corrosion depth, where the detection information is used to characterize the corrosion state of the concrete sample to be tested.

[0043] In this embodiment, the detector is moved at a constant speed to obtain response data in different time periods. The above response data can be understood as the electrical signal generated by the detector after characteristic processing of the gamma rays. The above response data is used to characterize the distribution changes of the chlorine element received by the detector in different time periods. Because the response data has time correlation, the signal characteristics received by the detector in different time periods are different. For example, the first time period response data obtained in the first time period can better reflect the distribution of the chlorine element and is used to determine the above concentration response data. The second time period response data obtained in the second time period can better reflect the information of the corrosion depth of the chlorine element and is used to determine the above depth response data.

[0044] Furthermore, the concentration response data and depth response data for corrosion state detection can be determined based on the response data of different time periods. The above-mentioned concentration response data is used to characterize the concentration of chlorine in concrete, and the above-mentioned depth response data is used to characterize the corrosion depth of concrete. Optionally, the concentration response data and depth response data can be determined by calculating the integral and the rate of change, and the interval integral of the response data of the first time period is used as the concentration response data, and the response change rate of the response data of the second time period is used as the depth response data. Optionally, the concentration response data and depth response data can be determined by feature extraction, and the peak value after normalization of the response data of the first time period is used as the concentration response data. The above-mentioned peak value can reflect the chlorine concentration as a key feature point, and the fitting slope of the response data of the second time period is used as the depth response data. The above-mentioned slope can reflect the changing trend of the current corrosion depth.

[0045] In this embodiment, the concentration response data and the depth response data reflect the corrosion of concrete from different dimensions. Since the concentration response data and the depth response data are determined based on the response data of different time periods, the depth response data will also reflect some chlorine element concentration information, and the concentration response data will also reflect some corrosion depth information. For example, please refer to Figure 3 In one embodiment, the interval integral of the response data of the first period is used as the concentration response data, and the response change rate of the response data of the second period is used as the depth response data. Figure 3 It reflects the three-dimensional relationship between chlorine concentration, corrosion depth and response data, among which, Figure 3The chlorine concentration in the sample is represented by the different chlorine contents shown in the colored strips. For concrete samples with the same corroded area size, different chlorine contents can be regarded as different chlorine concentrations.

[0046] Depend on Figure 3 It can be seen that there is a certain correlation between concentration response data and depth response data. For example, different data points of the same color represent different concrete samples with the same chlorine content. Each data point is in the same numerical dimension plane of the response integral and corresponds to a different corrosion depth. For another example, different data points on the numerical dimension plane of the same corrosion depth represent different concrete samples with the same corrosion depth. If the colors of each data point are different, it indicates that each data point has a different chlorine content, which is reflected in the different average change rates corresponding to each data point. When performing a single concentration test or depth test, sending both the concentration response data and the depth response data of the same concrete sample to the corresponding concentration detection model or depth detection model can better resist the noise influence of the single data, thereby improving the accuracy of the related detection.

[0047] Furthermore, in this embodiment, by setting different response weights for the concentration response data and the depth response data, the detection of the concrete sample to be tested based on the concentration response data and the depth response data can be made more focused on their respective detection directions. Specifically, when a detection instruction is triggered and issued by the host computer or the user, when a concentration detection is performed in response to the concentration detection instruction, the concentration response data and the depth response data are processed based on the first preset weight, and the chlorine concentration of the concrete sample to be tested is output. When a depth detection is performed in response to the depth detection instruction, the concentration response data and the depth response data are processed based on the second preset weight, and the corrosion depth of the concrete sample to be tested is output. By reasonably setting and optimizing the two weight coefficients, the prediction accuracy can be improved, ensuring that the chlorine concentration and corrosion depth of the concrete sample can be accurately predicted in practical applications.

[0048] In this embodiment, detection information characterizing the corrosion state is generated by detecting the chlorine concentration and corrosion depth. The above detection information is used to characterize the degree of corrosion of the concrete sample and help the detection personnel understand the health status of the concrete structure. For example, a lower chlorine concentration and a shallower corrosion depth may indicate that the concrete structure is relatively healthy, while a higher chlorine concentration and a deeper corrosion degree may indicate that the concrete structure has been severely corroded and needs to take corresponding maintenance or repair measures. Since the chlorine concentration and corrosion depth after the above model detection are more accurate, the detection personnel can formulate a more detailed detection standard according to industry specifications to better distinguish the corrosion conditions of the concrete structure according to the chlorine concentration and corrosion depth. Optionally, the chlorine concentration and corrosion depth can be fused through a specific algorithm or rule to generate a digitized maintenance information.

[0049] The technical solution provided by this embodiment dynamically analyzes response data from different time periods to obtain more accurate information describing the corrosion state of concrete caused by chlorine. Specifically, a mobile detector dynamically acquires response data from the concrete sample under test at different time periods, and then performs targeted detection on the response data from the first and second time periods to obtain the chlorine concentration and corrosion depth of the concrete sample under test as corrosion information, rather than simply detecting chlorine content as corrosion information. Compared with existing technologies, by dynamically acquiring and analyzing response data, deeper corrosion information, including chlorine concentration and corrosion depth, can be obtained, enabling accurate prediction of the chlorine corrosion state in concrete and obtaining more precise chlorine corrosion information.

[0050] In one embodiment, multiple detector response values ​​received during different time periods of the concrete sample to be tested are obtained as first-time period response data and second-time period response data. Specifically, a detector is used to approach the concrete sample to be tested at a constant speed, and multiple detector response values ​​received during different time periods during the constant approach are obtained. The multiple detector response values ​​of the detector during the first time period are used as the first-time period response data, and the multiple detector response values ​​of the detector during the second time period are used as the second-time period response data. The detector response values ​​can be understood as electrical signals converted after the detector captures gamma rays. The first time period can be a period in which the response data changes significantly under different corrosion depths. Therefore, using the first-time period response data as data reflecting the corrosion depth can more accurately detect the corrosion depth of the concrete sample to be tested. The second time period can be a period in which the response data changes significantly under different chlorine concentrations. Therefore, using the second-time period response data as data reflecting the chlorine concentration can more accurately detect the chlorine concentration of the concrete sample to be tested.

[0051] For example, the response time interval of 45s to 60s when the detector is close to the concrete sample can be selected as the first time period, and the response time interval of 27s to 40s when the detector is close to the concrete sample can be selected as the second time period. In the detection of concrete samples with different corrosion depths, the response data in the response time interval of 45s to 60s shows more obvious changes in values ​​under different corrosion depths than other time periods, so the response time interval of 45s to 60s is selected as the first time period. In the detection of concrete samples with different chlorine concentrations, the response data in the response time interval of 27s to 40s shows more obvious changes in values ​​under different chlorine concentrations than other time periods, so the response time interval of 27s to 40s is selected as the second time period.

[0052] In this implementation, the detector response values ​​for the concrete sample at different time periods are obtained as response data for different detection targets. By moving the detector toward the concrete sample at a constant speed and selecting the response data from the first and second time periods, where the data changes significantly for different detection targets, for targeted analysis, this ensures effective differentiation and accurate prediction of the chlorine concentration and corrosion depth of the concrete sample, thereby achieving a precise prediction of the chlorine corrosion state in the concrete.

[0053] In one embodiment, the interval integral of the response data of the first time period is used as the concentration response data, and the response change rate of the response data of the second time period is used as the depth response data. Specifically, the response integral of the response data of the first time period in the first time period is obtained, and the response integral is used as the concentration response data, and the response change rate of the response data of the second time period in the second time period is obtained, and the response change rate is used as the depth response data, wherein the response time interval represented by the second time period is earlier than the response time interval represented by the first time period. Preferably, the first time period is a response time interval of 45s to 60s, and the second time period is a response time interval of 27s to 40s.

[0054] Refer to Figures 4(a) and 4(b). Different colors in Figure 4(a) represent different chlorine concentrations, and different colors in Figure 4(b) represent different corrosion depths. Normalizing the response values ​​to achieve standardization facilitates observation and interpretation of the changing trends and characteristics of the response values. Figure 4(a) shows that when the detector approaches a concrete sample with a fixed corroded area size, the same corrosion depth, and different chlorine concentrations at a constant speed, the detector responds stably within the first period, using the response time interval of 45s to 60s as the first period (i.e., the response integral prediction interval in 4(a)). The response integral can distinguish smaller concentration differences. Therefore, the response integral of the first period response data can be used as the concentration response data. As shown in Figure 4(b), when the detector approaches a concrete sample with a fixed corroded area size, the same chlorine concentration, and different corrosion depths at a constant speed, the detector response varies significantly during the second period, as the response time interval of 27s to 40s is used as the second period (i.e., the average rate of change prediction interval in Figure 4(b)). Since the response rate of change can discern more spatial information about the chlorine element, the response rate of the second period can be used as the depth response data. It should be noted that in this example, the concrete samples were located in the same experimental environment, and except for the chlorine concentration or corrosion depth, all other sample and experimental parameters remained the same. In particular, the detector movement speed and the corroded area were kept the same, for example, the detector movement speed was kept constant at 1 cm / s, and the corroded area size was 50 cm × 50 cm × 3 cm.

[0055] Exemplarily, the above concentration response data can be expressed as ,in, is the response data for the first period, is the response time of 45 seconds, is the response time of the 60th second, and the above depth response data can be expressed as , in, The response data for the second period is is the response time of 27 seconds, The response time is 40 seconds.

[0056] For example, in one embodiment, referring to Figures 5(a) and 5(b), Figure 5(a) illustrates an example detection scenario at a certain point in time within the response time interval of 27s to 40s when detector 21 approaches a concrete sample. Figure 5(b) illustrates an example detection scenario at a certain point in time within the response time interval of 45s to 60s when detector 21 approaches a concrete sample 23. The dashed box represents the scanning range of detector 21. As detector 21 approaches corroded region 22, the detectable reaction zone 24 becomes increasingly larger. The response time interval of 27s to 40s is more suitable for detecting the corrosion depth of concrete sample 23, while the response time interval of 45s to 60s is more suitable for detecting the chlorine concentration of concrete sample 23.

[0057] In this implementation, concentration response data and depth response data are determined based on the detector response characteristics at different time intervals. By leveraging the response integral's ability to distinguish small concentration differences, the response data for the first time interval is adjusted to concentration response data. Furthermore, by leveraging the response rate of change's ability to capture more spatial information about chlorine, the response data for the second time interval is adjusted to depth response data. This effectively improves the accuracy and effectiveness of detecting chlorine concentration and corrosion depth in concrete samples, enabling precise prediction of the chlorine corrosion state in concrete.

[0058] In one embodiment, a concentration detection model adjusts and analyzes the concentration response data and the depth response data based on a first preset weight, and the concentration detection model outputs the chlorine concentration of the concrete sample to be tested. Specifically, the concentration detection model generates concentration mapping parameters based on the concentration response data and the depth response data based on the first preset weight, wherein the first preset weight is used to represent the concentration detection model's attention to the concentration response data. The concentration mapping parameters are mapped to a high-dimensional concentration space to obtain a corresponding high-dimensional concentration mapping vector. The chlorine concentration of the concrete sample to be tested is generated based on the high-dimensional concentration mapping vector and a concentration correlation function.

[0059] The above-mentioned concentration mapping parameters can be understood as the relevant features that characterize the concentration of chlorine in the concrete sample to be tested. The concentration response data and the depth response data are weighted and fused based on the first preset weight to obtain the concentration mapping parameters, so that the concentration detection model can highlight the prediction of chlorine concentration while considering the potential impact of the depth response data on the concentration prediction. The above-mentioned concentration mapping parameters can be expressed as ,in, is the first preset weight, is the concentration mapping parameter.

[0060] The above-mentioned concentration high-dimensional space can be understood as a high-dimensional feature space with chlorine element concentration characteristics after mapping through the kernel function. The concentration mapping parameters are implicitly mapped to the concentration high-dimensional space through the kernel function, making it easier to find decision boundaries or regression relationships that can distinguish different chlorine element concentrations in the high-dimensional space, which helps to improve the accuracy of the concentration detection model in predicting chlorine element concentration. The above-mentioned high-dimensional concentration mapping vector can be understood as the representation of the concentration mapping parameters in the concentration high-dimensional space. For example, the kernel function is used to map each concentration mapping parameter to a point in an infinite-dimensional space, and this point is represented as a high-dimensional concentration mapping vector, so as to better reflect the complex relationship of chlorine element concentration. The above-mentioned concentration-related function can be understood as a function that converts a high-dimensional concentration mapping vector into a predicted value of chlorine element concentration. The above-mentioned concentration-related function can be expressed as The concentration correlation function is determined through the model training process, which maps the high-dimensional concentration mapping vector to the corresponding chlorine element concentration as the predicted value based on the rules learned from the training data.

[0061] In this embodiment, a concentration detection model is used to detect the chlorine concentration of a concrete sample based on concentration response data and depth response data. Specifically, the concentration response data and depth response data are processed according to first preset weights to generate a high-dimensional concentration mapping vector. This high-dimensional concentration mapping vector is then converted into a predicted chlorine concentration value for the concrete sample using a concentration-correlation function. By rationally utilizing weights, feature mapping, and function transformation, this embodiment achieves accurate prediction of chlorine concentration.

[0062] In one embodiment, the depth detection model adjusts and analyzes the concentration response data and the depth response data based on a second preset weight, and the depth detection model outputs the corrosion depth of the concrete sample to be tested. Specifically, the depth detection model generates depth mapping parameters based on the concentration response data and the depth response data based on the second preset weight, where the second preset weight is used to represent the depth detection model's attention to the depth response data. The depth mapping parameters are mapped into a deep high-dimensional space to obtain a corresponding high-dimensional depth mapping vector. The corrosion depth of the concrete sample to be tested is then generated based on the high-dimensional depth mapping vector and a depth correlation function.

[0063] The above-mentioned depth mapping parameters can be understood as the relevant features that characterize the depth of chlorine corrosion in concrete samples. The concentration response data and the depth response data are weighted and fused based on the second preset weight to obtain the depth mapping parameters, so that the depth detection model can highlight the prediction of the depth of chlorine corrosion while considering the potential impact of the concentration response data on the depth prediction. The above-mentioned depth mapping parameters can be expressed as ,in, is the second preset weight, is the depth map parameter.

[0064] The above-mentioned deep high-dimensional space can be understood as a high-dimensional feature space with chlorine corrosion depth characteristics after mapping through a kernel function. The depth mapping parameters are implicitly mapped to the deep high-dimensional space through the kernel function, making it easier to find decision boundaries or regression relationships that can distinguish different chlorine corrosion depths in the deep high-dimensional space, which helps to improve the accuracy of the depth detection model in predicting corrosion depth. The above-mentioned high-dimensional depth mapping vector can be understood as the representation of the depth mapping parameters in the deep high-dimensional space. For example, a kernel function is used to map each depth mapping parameter to a point in an infinite-dimensional space, and this point is represented as a high-dimensional depth mapping vector, thereby better reflecting the complex relationship of the chlorine corrosion depth. The above-mentioned depth correlation function can be understood as a function that converts a high-dimensional depth mapping vector into a corrosion depth prediction value. The above-mentioned depth correlation function can be expressed as The depth correlation function is determined through the model training process, which maps the high-dimensional depth map vector to the corresponding chlorine corrosion depth as the predicted value based on the rules learned from the training data.

[0065] In this embodiment, a depth detection model is used to detect the chlorine corrosion depth of a concrete sample based on concentration response data and depth response data. Specifically, the concentration response data and depth response data are processed according to a second preset weight to generate a high-dimensional depth map vector. This high-dimensional depth map vector is then converted into a predicted chlorine corrosion depth value for the concrete sample using a depth correlation function. By rationally utilizing weights, feature mapping, and function conversion, this embodiment achieves accurate prediction of chlorine corrosion depth.

[0066] In a possible implementation, the concentration detection model and the depth detection model are pre-trained SVR models, wherein the concentration detection model can be trained according to the following steps S61 to S65:

[0067] S61: Acquire concentration response data and depth response data of multiple groups of concrete training samples, and acquire a preset concentration and preset depth of chlorine element in each group of concrete simulation samples, wherein the preset concentration and preset depth of chlorine element in each group of concrete simulation samples are different.

[0068] S63: Generate concentration mapping parameters using the first preset weights according to the concentration response data and the depth response data of each group of concrete training samples.

[0069] S65: generating a concentration-related function of the concentration detection model according to the concentration mapping parameters and the preset concentration of each group of concrete training samples, and training the concentration detection model using a first SVR loss function.

[0070] The above-mentioned concentration detection model can be trained according to the following steps S71 to S75:

[0071] S71: Acquire concentration response data and depth response data of multiple groups of concrete training samples, and acquire a preset concentration and preset depth of chlorine element in each group of concrete simulation samples, wherein the preset concentration and preset depth of chlorine element in each group of concrete simulation samples are different.

[0072] S73: Generate depth mapping parameters using the second preset weights according to the concentration response data and the depth response data of each group of concrete training samples.

[0073] S75: generating a depth-related function of the depth detection model according to the depth mapping parameters of each group of concrete training samples and the preset depth, and training the depth detection model using a second SVR loss function.

[0074] In this embodiment, training is performed by obtaining concrete training samples with varying chlorine concentrations and corrosion depths, ensuring that the model can learn corrosion characteristics under different conditions during training. Given the actual chlorine concentration and corrosion depth of any set of concrete training samples, the actual chlorine concentration of the concrete training samples is used as a preset concentration, and the actual chlorine corrosion depth of the concrete training samples is used as a preset depth. These preset concentrations and depths can serve as standard values, enabling the concentration detection model and depth detection model to learn the complex mapping relationship between response data and preset concentrations or preset depths. Simultaneously, the concentration response data and depth response data are weighted and summed according to a first preset weight or a second preset weight to generate concentration mapping parameters or depth mapping parameters as training sample features.

[0075] Further, please refer to Figures 6(a) and 6(b). Data points of different colors represent concrete samples with different chlorine concentrations. Figure 6(a) shows the two-dimensional feature mapping relationship between the concentration response data, corrosion depth, and chlorine concentration in the first period of one example. In concrete samples with the same corrosion depth, the response integral values ​​of the concrete samples within 45s to 60s are different due to different chlorine contents. The concentration detection model generates a concentration-related function by learning the mapping relationship between the concentration mapping parameters and preset concentrations of different samples. Figure 6(b) shows the two-dimensional feature mapping relationship between the depth response data, corrosion depth, and chlorine concentration in the second period of one example. In concrete samples with the same chlorine content, the average response change rate of the concrete samples within 27s to 40s is different due to different corrosion depths. The depth detection model generates a depth-related function by learning the mapping relationship between the depth mapping parameters and preset depths of different samples, allowing the model to learn a wider range of corrosion characteristics, thereby enhancing the model's adaptability and generalization ability for different types of concrete samples, and thus more accurately predicting the chlorine concentration and corrosion depth of concrete.

[0076] In this embodiment, the SVR loss function is used for model training to improve the prediction accuracy. Specifically, the first SVR loss function is used to train the concentration detection model, and the second SVR loss function is used to train the depth detection model. During the initial training process, the first SVR loss function and the second SVR loss function can be expressed as , in, The contribution weight of each dimension to the corrosion depth is, and is the slack variable, is a regularization parameter used to control the trade-off between model complexity and training error. is the kernel function mapping, is the overall deviation of the response data, is the maximum lossless deviation between the predicted value and the true value.

[0077] In one embodiment, a program can simulate scenarios involving concrete samples with varying chlorine concentrations and corrosion depths. Specifically, a particle transport model is established to simulate different chlorine concentrations and corrosion depths in concrete. This particle transport model can be understood as a model capable of simulating the interaction between neutrons and chlorine in concrete, as well as the generation of prompt gamma rays for propagation and detection. By invoking a set program to automatically run the simulation in batches, a large amount of detector response data and preset data can be obtained under different chlorine concentrations and corrosion depths, which can be used for training.

[0078] For example, in the above simulation process, the chlorine concentration and corrosion depth are evenly distributed at different depths and positions, and the corrosion area is set to a size of 50cm×50cm×3cm to simulate the large-area corrosion situation that may be encountered in actual detection. By setting the detector to approach the concrete sample at a uniform speed of 1cm / s, the sample detection process at different depths and different chlorine concentrations is simulated, and the detector response data is recorded within 5s to 60s, thereby obtaining a more comprehensive and diverse training data set, and being able to obtain a large amount of high-quality training data in a relatively short time.

[0079] In this implementation, an SVR model is used to construct a concentration detection model and a depth detection model to accurately predict the chlorine concentration and corrosion depth of concrete. By training multiple sets of samples with different conditions, the concentration detection model and the depth detection model learn the complex mapping relationship between the response data and the chlorine concentration and corrosion depth, enhancing the adaptability of corrosion state detection for concrete under different corrosion states. At the same time, the SVR loss function is used to balance model complexity and training error, further improving the robustness and accuracy of the model, thereby improving the prediction accuracy of concrete chlorine concentration and corrosion depth, and achieving accurate prediction of the chlorine corrosion state in concrete.

[0080] In one possible implementation, after step S65, the trained concentration detection model may be optimized, and after step S75, the trained depth detection model may be optimized. Evaluation indicators and pre-set conditions provide quantitative criteria for model verification and optimization, ensuring that the model can accurately assess the chlorine concentration and corrosion depth in concrete, thereby ensuring accurate prediction of the chlorine corrosion state in concrete in practical applications. Optimizing the trained concentration detection model may include the following steps:

[0081] S601: Acquire multiple groups of concrete prediction samples, obtain the actual concentration of chlorine in any concrete prediction sample, and generate the predicted concentration of chlorine in the concrete prediction sample based on the concentration detection model.

[0082] S603: Generate a prediction error and an evaluation index of the concentration detection model based on the true concentration and the predicted concentration. If the evaluation index does not meet the preset conditions, repeatedly adjust the first preset weight and the first SVR loss function according to the prediction error until the evaluation index of the concentration detection model meets the preset conditions.

[0083] In this embodiment, optimizing the trained depth detection model may include the following steps:

[0084] S701: Acquire multiple groups of concrete prediction samples, obtain the true depth of chlorine in any concrete prediction sample, and generate the predicted depth of chlorine in the concrete prediction sample based on the concentration detection model.

[0085] S703: Generate a prediction error and an evaluation index of the depth detection model based on the true depth and the predicted depth. If the evaluation index does not meet the preset conditions, repeatedly adjust the second preset weight and the second SVR loss function according to the prediction error until the evaluation index of the depth detection model meets the preset conditions.

[0086] In this embodiment, the concentration detection model is optimized based on the actual concentration and predicted concentration of the concrete prediction sample. Specifically, the actual concentration of chlorine in any concrete prediction sample and the predicted concentration of chlorine predicted by the concentration detection model are obtained, and the difference between the actual concentration and the predicted concentration is used as the prediction error. A series of evaluation indicators, such as the coefficient of determination, mean absolute error, and root mean square error, are calculated based on the actual depth and the predicted depth. If the evaluation indicators do not meet the preset conditions, the first preset weight and the first SVR loss function are adjusted according to the prediction error. The above prediction error can be expressed as , adjust the first preset weight and the first SVR loss function according to the following method: , , is the first preset weight after adjustment, is the adjusted first SVR loss function.

[0087] In this embodiment, the depth detection model is optimized based on the actual depth and predicted depth of the concrete prediction sample. Specifically, the actual corrosion depth of the chlorine element in any concrete prediction sample and the predicted corrosion depth of the chlorine element predicted by the depth detection model are obtained, and the difference between the actual depth and the predicted depth is used as the prediction error. Based on the actual depth and the predicted depth, a series of evaluation indicators are calculated, such as the coefficient of determination, the mean absolute error, and the root mean square error. If the evaluation indicators do not meet the preset conditions, the second preset weight and the second SVR loss function are adjusted according to the prediction error. Among them, the above prediction error can be expressed as , adjust the second preset weight and the second SVR loss function according to the following method: , , is the second preset weight after adjustment, is the adjusted second SVR loss function.

[0088] In this embodiment, the above evaluation indicators include the coefficient of determination, mean absolute error and root mean square error. , ,in, is the coefficient of determination, is the prediction error, is the true value, is the predicted value, is the mean of the true values, The number of concrete samples predicted. The above true value is the true concentration or true depth, and the above predicted value is the predicted concentration or predicted depth. The linear relationship between the true value and the predicted value is used as the determination coefficient to measure the fitting effect of the model. The closer it is to 1, the better the model fit is. ,in, It is used to measure the mean absolute error between the predicted value and the true value. The smaller the value of, the higher the prediction accuracy of the model. ,in, The smaller the value of , the smaller the model error and the higher the prediction accuracy.

[0089] Optionally, the above-mentioned preset condition may be the number of rounds of model adjustment represented by the evaluation index. When the number of model adjustment rounds is less than or equal to the preset number of times, the evaluation index is deemed to not meet the preset condition. At this time, the preset weights and loss functions of the corresponding model are readjusted according to the prediction error to form a new concentration prediction model and depth prediction model, and the evaluation indexes of the new concentration prediction model and depth prediction model are re-judged whether they meet the preset conditions. When the number of model adjustment rounds is greater than the preset number of times, the evaluation index is deemed to meet the preset conditions, and the concentration prediction model and depth prediction model that have undergone the preset number of iterations are used as the final prediction models.

[0090] Preferably, the above-mentioned preset conditions can be a coefficient of determination, mean absolute error, and root mean square error with respective preset thresholds. When any of the coefficient of determination, mean absolute error, and root mean square error of the concentration detection model or the depth detection model does not reach the preset threshold of the respective indicator, the preset weight and loss function of the corresponding detection model are adjusted. When the coefficient of determination, mean absolute error, and root mean square error of the corresponding concentration detection model or the depth detection model all reach the preset threshold, the preset weight and loss function corresponding to the model are no longer adjusted. For example, when the coefficient of determination is greater than 0.95, the mean absolute error is less than 0.05%, and the root mean square error is less than 0.08%, the preset weight and loss function of the corresponding model are no longer adjusted.

[0091] For example, please refer to Figure 7(a). Figure 7(a) shows the detection of chlorine concentration of concrete prediction samples by the trained concentration detection model. The predicted value in Figure 7(a) represents the predicted concentration. The ideal line in Figure 7(a) can be understood as the standard line representing the linear relationship of the true concentration. By obtaining the prediction error between the predicted concentration and the true concentration, the determination coefficient, mean absolute error and root mean square error of the concentration detection model can be calculated, and then the concentration detection model can be continuously optimized based on the prediction error and evaluation indicators.

[0092] For example, please refer to Figure 7(b), which shows the detection of the chlorine corrosion depth of the concrete prediction sample by the trained depth detection model. The predicted value in Figure 7(b) represents the predicted depth. The ideal line in Figure 7(b) can be understood as the standard line representing the linear relationship of the true depth. By obtaining the prediction error between the predicted depth and the true depth, the determination coefficient, mean absolute error and root mean square error of the depth detection model can be calculated, and then the depth detection model can be continuously optimized according to the prediction error and evaluation indicators.

[0093] See also Figure 8 The present application also provides a device for detecting the corrosion state of concrete, the device comprising:

[0094] The data acquisition unit 100 is configured to acquire response data of a first time period and response data of a second time period of the concrete sample to be tested, determine concentration response data based on the response data of the first time period, and determine depth response data based on the response data of the second time period;

[0095] The data processing unit 200 is configured to, in response to a concentration detection instruction, input the concentration response data and the depth response data into a concentration detection model, wherein the concentration detection model outputs the chlorine concentration of the concrete sample to be tested based on a first preset weight; and, in response to a depth detection instruction, input the concentration response data and the depth response data into a depth detection model, wherein the depth detection model outputs the corrosion depth of the concrete sample to be tested based on a second preset weight;

[0096] The information generating unit 300 is configured to generate detection information based on the chlorine concentration and the corrosion depth, wherein the detection information is used to characterize the corrosion state of the concrete sample to be tested.

[0097] in,

[0098] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0099] In an embodiment of the present application, a device for detecting a concrete corrosion state is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, or other devices that can provide the above functions.

[0100] On the other hand, the present application further provides a computer device, comprising: a memory storing computer instructions; and at least one processor configured to execute the computer instructions in the memory to perform the concrete corrosion state detection method in the above embodiment.

[0101] See also Figure 9 , Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. Figure 9 As shown, the computer device includes: one or more processors 101, memory 201, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of a GUI on an external input / output device (such as, a display device coupled to an interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 9 A processor 101 is taken as an example.

[0102] Processor 101 may be a central processing unit, a network processor, or a combination thereof. Processor 101 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0103] The memory 201 stores instructions that can be executed by at least one processor 101, so that the at least one processor 101 executes the method shown in the above embodiment.

[0104] The memory 201 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 201 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 201 may optionally include a memory remotely located relative to the processor 101, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0105] The memory 201 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 201 may also include a combination of the above types of memory.

[0106] The computer device further includes a communication interface 301 for the computer device to communicate with other devices or a communication network.

[0107] The embodiments of the present application also provide a computer-readable storage medium. The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0108] The present application is described with reference to the flowcharts and / or block diagrams of the methods and apparatuses according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0109] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0111] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0112] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the device embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0113] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

[0114] Although the embodiments of the present application have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations shall fall within the scope defined by the appended claims.

Claims

1. A method for detecting concrete corrosion status, characterized in that: The method comprises: Using a detector to obtain response data of a first period and response data of a second period of time of the concrete sample to be tested, determining concentration response data based on the response data of the first period of time, and determining depth response data based on the response data of the second period of time; In response to a concentration detection instruction, generating a chlorine concentration of the concrete sample to be tested according to the concentration response data and the depth response data under a first preset weight; In response to a depth detection instruction, generating a corrosion depth of the concrete sample to be tested according to the concentration response data and the depth response data under a second preset weight; generating detection information based on the chlorine concentration and the corrosion depth, wherein the detection information is used to characterize the corrosion state of the concrete sample to be tested; wherein, obtaining a response integral of the response data of the first time period in the first time period, and using the response integral as the concentration response data; A response change rate of the second-period response data in the second period is acquired, and the response change rate is used as the depth response data.

2. The method according to claim 1, characterized in that Acquiring the response data of the first period and the response data of the second period of the concrete sample to be tested includes: Using a detector to approach the concrete sample to be tested at a constant speed, and obtaining a plurality of detector response values ​​received by the detector at different time periods during the process of the detector approaching at a constant speed; The plurality of detector response values ​​of the detector in the first time period are used as the first time period response data, and the plurality of detector response values ​​of the detector in the second time period are used as the second time period response data.

3. The method according to claim 1 or 2, characterized in that The response time interval represented by the second time period is earlier than the response time interval represented by the first time period.

4. The method according to claim 1, wherein The chlorine concentration of the concrete sample to be tested is generated by a concentration detection model. The concentration detection model generates the chlorine concentration of the concrete sample to be tested in the following manner: The concentration detection model generates a concentration mapping parameter according to the concentration response data and the depth response data based on the first preset weight, wherein the first preset weight is used to represent the attention of the concentration detection model to the concentration response data; The concentration mapping parameter is mapped to a concentration high-dimensional space to obtain a corresponding high-dimensional concentration mapping vector, and the chlorine element concentration of the concrete sample to be tested is generated according to the high-dimensional concentration mapping vector and a concentration correlation function.

5. The method according to claim 1, wherein The corrosion depth of the concrete sample to be tested is generated by a depth detection model. The depth detection model generates the corrosion depth of the concrete sample to be tested in the following manner: The depth detection model generates a depth mapping parameter according to the concentration response data and the depth response data based on the second preset weight, wherein the second preset weight is used to represent the attention of the depth detection model to the depth response data; The depth mapping parameters are mapped to a deep high-dimensional space to obtain a corresponding high-dimensional depth mapping vector, and the corrosion depth of the concrete sample to be tested is generated according to the high-dimensional depth mapping vector and a depth correlation function.

6. The method according to claim 4, characterized in that The concentration detection model is a pre-trained SVR model; the concentration detection model is trained in the following manner: Acquiring concentration response data and depth response data for multiple groups of concrete training samples, and obtaining a preset concentration and a preset depth of chlorine in each group of concrete simulation samples, wherein the preset concentration and the preset depth of chlorine in each group of concrete simulation samples are different; generating concentration mapping parameters using the first preset weights according to the concentration response data and the depth response data of each group of concrete training samples; A concentration-related function of the concentration detection model is generated according to the concentration mapping parameters and the preset concentration of each group of concrete training samples, and the concentration detection model is trained using a first SVR loss function.

7. The method according to claim 5, characterized in that The depth detection model is a pre-trained SVR model; the depth detection model is trained in the following manner: Acquiring concentration response data and depth response data for multiple groups of concrete training samples, and obtaining a preset concentration and a preset depth of chlorine in each group of concrete simulation samples, wherein the preset concentration and the preset depth of chlorine in each group of concrete simulation samples are different; generating depth mapping parameters using the second preset weights according to the concentration response data and the depth response data of each group of concrete training samples; A depth-related function of the depth detection model is generated according to the depth mapping parameters and the preset depth of each group of concrete training samples, and the depth detection model is trained using a second SVR loss function.

8. The method according to claim 6, characterized in that After training the concentration detection model, the method further includes: Acquire multiple groups of concrete prediction samples, obtain the actual concentration of chlorine in any concrete prediction sample, and generate a predicted concentration of chlorine in the concrete prediction sample based on the concentration detection model; A prediction error and an evaluation index of the concentration detection model are generated based on the true concentration and the predicted concentration. If the evaluation index does not meet the preset conditions, the first preset weight and the first SVR loss function are repeatedly adjusted according to the prediction error until the evaluation index of the concentration detection model meets the preset conditions.

9. The method according to claim 7, characterized in that After training the depth detection model, the method further includes: Acquire multiple groups of concrete prediction samples, obtain the true depth of chlorine in any concrete prediction sample, and generate the predicted depth of chlorine in the concrete prediction sample based on the concentration detection model; Based on the true depth and the predicted depth, a prediction error and an evaluation index of the depth detection model are generated. If the evaluation index does not meet the preset conditions, the second preset weight and the second SVR loss function are repeatedly adjusted according to the prediction error until the evaluation index of the depth detection model meets the preset conditions.

10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the concrete corrosion state detection method according to any one of claims 1 to 9 by executing the computer instructions.

Images