Methods, electronic equipment and systems for detecting internal media in steel structure columns

Abstract

This invention provides a method, electronic device, and system for detecting the internal medium of a steel structure column, relating to the field of ultrasonic testing technology. The method includes: acquiring a first echo signal from inside the target steel structure column; wherein the first echo signal is a high-frequency signal; determining whether clutter exists in the first echo signal; if so, acquiring a second echo signal from inside the target steel structure column; wherein the second echo signal is a low-frequency signal; adjusting the gain of the second echo signal based on a first preset gain adjustment method to obtain a recognition feature map; and obtaining the medium detection result inside the target steel structure column based on the recognition feature map. This invention improves the accuracy of detection, and the method is applicable to the detection of ice, water, and sand inside target steel structure columns, saving detection costs.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic testing technology, and in particular to a method, electronic device and system for detecting the internal medium of a steel structure column. Background Technology

[0002] In the field of steel structure engineering, columns are critical load-bearing components, and their safety is of paramount importance. However, steel structure columns face numerous problems caused by the internal environment during use.

[0003] On the one hand, in cold environments, water accumulating inside steel structure columns easily freezes and solidifies. The expansion of the ice puts pressure on the internal structure of the column, causing deformation and cracking, thus reducing its load-bearing capacity. In severe cases, it may even lead to collapse, posing a significant threat to the safety of pedestrians and vehicles. On the other hand, during daily use, dust may enter the interior of the columns, and internal water accumulation can cause corrosion of the inner walls, resulting in the accumulation of iron filings. These factors accelerate the corrosion rate of the steel structure, increasing safety hazards and potentially leading to accidents.

[0004] Currently, there is a lack of effective, accurate, and reliable testing methods and processes for detecting the internal media of steel structure columns. Existing testing technologies have many limitations:

[0005] In the detection of icing and voiding inside steel structure columns, existing methods mainly include microwave testing and vibration testing. Microwave testing involves complex circuit design, requiring advanced equipment and technology. Furthermore, the metal casing of the steel structure can interfere with the microwave signal, severely affecting the accuracy of the test. Vibration testing requires precise installation and calibration of vibration sensors; however, external environmental vibrations and noise can easily interfere with the test results, necessitating cumbersome signal filtering and processing. Therefore, there is currently no ideal method for detecting icing and voiding inside steel structure columns.

[0006] In the detection of water mixtures, existing methods for detecting water accumulation inside steel structure columns mostly employ single-probe detection. However, when foreign objects are present at the bottom of the water, this method cannot accurately detect the internal conditions, making it difficult to meet practical needs. Summary of the Invention

[0007] This invention provides a method, electronic device, and system for detecting the internal medium of steel structure columns, in order to solve the problem of inaccurate current methods for detecting the internal medium of steel structure columns.

[0008] In a first aspect, embodiments of the present invention provide a method for detecting the internal medium of a steel structure column, comprising:

[0009] The first echo signal inside the target steel structure column was acquired; the first echo signal was a high-frequency signal.

[0010] Determine whether there is noise in the first echo signal;

[0011] If present, the second echo signal inside the target steel structure column is acquired; the second echo signal is a low-frequency signal.

[0012] Based on the first preset gain adjustment method, the gain of the second echo signal is adjusted to obtain the identification feature map;

[0013] Based on the identified feature map, the detection results of the medium inside the target steel structure column are obtained.

[0014] In one possible implementation, determining whether clutter exists in the first echo signal includes:

[0015] If there is no water accumulation echo in the first echo signal, the gain of the first echo signal is adjusted based on the second preset gain adjustment method.

[0016] Determine whether there is noise in the first echo signal after gain adjustment.

[0017] In one possible implementation, the gain of the second echo signal is adjusted based on a first preset gain adjustment method to obtain a recognition feature map, including:

[0018] Within a preset gain range, the gain of the second echo signal is adjusted according to a preset sequence to obtain the second echo signal after gain adjustment.

[0019] The identification feature map is obtained based on the second echo signal after gain adjustment.

[0020] In one possible implementation, or, based on a first preset gain adjustment method, the gain of the second echo signal is adjusted to obtain the identification feature map, the method further includes:

[0021] Based on a preset gain value, the gain of the second echo signal is adjusted to obtain the second echo signal after gain adjustment.

[0022] The identification feature map is obtained based on the second echo signal after gain adjustment.

[0023] In one possible implementation, the medium detection results inside the target steel structure column are obtained based on the identified feature map, including:

[0024] The recognition feature map is input into the pre-trained first target detection model to obtain the medium detection result inside the target steel structure column;

[0025] The test results for the medium inside the target steel structure column include one of the following: water, air, sand, ice, ice-water mixture, and water-sand mixture.

[0026] In one possible implementation, if the medium detection result inside the target steel structure column is ice or an ice-water mixture, the method further includes:

[0027] The recognition feature map is input into a pre-trained second target detection model to determine the thickness of the ice detachment.

[0028] In one possible implementation, if the medium detection result inside the target steel structure column is sand or a mixture of sand and water, the method further includes:

[0029] Obtain the propagation speed of the second echo signal within the target steel structure column;

[0030] The target calculation formula is determined based on the propagation speed;

[0031] The thickness of the sand and gravel is determined based on the propagation speed and target calculation formula.

[0032] In one possible implementation, the medium detected inside the target steel structure column is a water-sand mixture, and the method further includes:

[0033] Time-frequency analysis was performed on the second echo signal after gain adjustment to identify the coherent peaks in the second echo signal.

[0034] Determine the characteristic parameters corresponding to the coherent peak, including frequency, duration, and amplitude;

[0035] Based on the characteristic parameters, the propagation speed of the second echo signal in the water-sand mixture and the thickness of the sand are determined.

[0036] In a second aspect, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0037] Thirdly, embodiments of the present invention provide a system including a high-frequency module, a low-frequency module, and an electronic device as provided in the second aspect; wherein the high-frequency module includes a high-frequency single probe, and the low-frequency module includes a low-frequency dual probe.

[0038] In this embodiment of the invention, a method combining high-frequency and low-frequency signals is employed for judgment. This multi-dimensional information fusion overcomes the limitations of single technologies, balancing detection depth and accuracy while enhancing anti-interference capabilities and environmental adaptability. First, based on the first echo signal, an initial measurement of the medium inside the target steel structure column is performed. The presence of water echoes in the first echo signal indicates the presence of water inside the target steel structure column. The absence of water echoes suggests the possible presence of a mixture or air, thus requiring further judgment of the first echo signal. Further judgment reveals the presence of noise in the first echo signal, increasing the probability of a mixture inside the target steel structure column. To determine the specific type of mixture, i.e., the medium, a second echo signal is used to further identify the possible types of mixtures. When the gain is too low, weak signals may be ignored; when the gain is too high, noise interference will be excessively amplified. Therefore, by adjusting the gain of the second echo signal, a characteristic echo reflecting the medium's characteristics is obtained. This characteristic echo is used as a recognition feature map for detection, ultimately determining the type of medium inside the target steel structure column. This method improves the accuracy of detection and is applicable to the detection of ice, water, and sand inside target steel structure columns, saving detection costs. Attached Figure Description

[0039] Figure 1 This is a system architecture diagram provided in an embodiment of the present invention;

[0040] Figure 2 This is a flowchart illustrating the implementation of the method for detecting the internal medium of a steel structure column provided in this embodiment of the invention.

[0041] Figure 3a This is a schematic diagram of the first echo signal provided by an embodiment of the present invention when the internal medium of the steel structure column is air;

[0042] Figure 3b This is a schematic diagram of the first echo signal when the internal medium of the steel structure column is water, provided by an embodiment of the present invention.

[0043] Figure 3c This is a schematic diagram of the first echo signal provided by an embodiment of the present invention when the internal media of a steel structure column are sand and gravel;

[0044] Figure 3d This is a schematic diagram of the first echo signal provided by an embodiment of the present invention when the internal medium of the steel structure column is a mixture of water and sand;

[0045] Figure 3e This is a schematic diagram of the first echo signal provided by an embodiment of the present invention when the internal medium of the steel structure column is ice;

[0046] Figure 3fThis is a schematic diagram of the first echo signal corresponding to the water-sand mixture after adjustment by the second preset gain adjustment method provided in this embodiment of the invention;

[0047] Figure 4a This is a schematic diagram of the second echo signal of ice at a gain of 64dB provided in an embodiment of the present invention;

[0048] Figure 4b This is a schematic diagram of the second echo signal of ice at a gain of 47dB provided in an embodiment of the present invention;

[0049] Figure 5a This is a water accumulation identification feature map provided in an embodiment of the present invention;

[0050] Figure 5b This is an ice identification feature map provided in an embodiment of the present invention;

[0051] Figure 5c This is a feature map of sand and gravel provided in an embodiment of the present invention;

[0052] Figure 5d This is an identification feature diagram of a water-sand mixture provided in an embodiment of the present invention;

[0053] Figure 5e This is an air identification feature map provided in an embodiment of the present invention;

[0054] Figure 5f This is an identification feature diagram of an ice-water mixture provided in an embodiment of the present invention;

[0055] Figure 6a This is an identification feature map of ice under 8mm delamination provided in an embodiment of the present invention;

[0056] Figure 6b This is an identification feature diagram of ice under 15mm delamination provided in an embodiment of the present invention;

[0057] Figure 6c This is an identification feature diagram of ice under 20mm delamination provided in an embodiment of the present invention;

[0058] Figure 6d This is an identification feature diagram of ice under 30mm void provided in an embodiment of the present invention;

[0059] Figure 6e This is an identification feature diagram of ice under 40mm delamination provided in an embodiment of the present invention;

[0060] Figure 6f This is an identification feature diagram of an ice-water mixture under 40mm de-cavitation provided in an embodiment of the present invention;

[0061] Figure 7a This is an identification feature diagram of 5mm thick sand and gravel provided in an embodiment of the present invention;

[0062] Figure 7b This is an identification feature diagram of 10mm thick sand and gravel provided in an embodiment of the present invention;

[0063] Figure 7c This is an identification feature diagram of 20mm thick sand and gravel provided in an embodiment of the present invention;

[0064] Figure 7d This is an identification feature diagram of 25mm thick sand and gravel provided in an embodiment of the present invention;

[0065] Figure 7e This is an identification feature diagram of 30mm thick sand and gravel provided in an embodiment of the present invention;

[0066] Figure 7f This is an identification feature diagram of 40mm thick sand and gravel provided in an embodiment of the present invention. Detailed Implementation

[0067] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0068] Figure 1 This is a system architecture diagram provided in an embodiment of the present invention; as shown below. Figure 1 As shown, this system is an ultrasonic testing system, which includes a high-frequency module 1, a low-frequency module 2, and electronic equipment 3. The high-frequency module 1 includes a high-frequency single probe 10 for acquiring the first echo signal; the low-frequency module 2 includes a low-frequency dual probe 20 for acquiring the second echo signal. The high-frequency single probe 10 can be 2.5MHz, and the low-frequency dual probe 20 can be 50kHz. The low-frequency dual probe 20 uses a one-transmit, one-receive mode for testing.

[0069] The electronic device 3 in the system includes a processor 30 and a memory 31. The memory 31 stores a computer program 32. When the processor 30 executes the computer program 32, it is used to implement a method for detecting the internal medium of a steel structure column.

[0070] Figure 2 This is a flowchart illustrating the implementation of the method for detecting the internal medium of a steel structure column provided in this embodiment of the invention. The following is a summary of the process. Figure 2 This paper describes a method for detecting the internal medium of a steel structure column, such as... Figure 2 As shown, the method may include:

[0071] Step 110: Acquire the first echo signal inside the target steel structure column; wherein, the first echo signal is a high-frequency signal.

[0072] In this embodiment, a high-frequency signal can be acquired by a 2.5MHz high-frequency single probe 10 in the high-frequency module 1 as the first echo signal, and the acquired first echo signal can be transmitted to the electronic device 3.

[0073] Step 120: Determine whether there is noise in the first echo signal.

[0074] Because high-frequency signals have weak penetration and are easily interfered with, analysis of high-frequency signals reveals that the presence of ice, sand, or other debris will generate interference in the first echo signal, creating clutter. Therefore, this embodiment uses the presence of clutter in the first echo signal as the criterion for determining whether subsequent detection is needed, thus reducing computational load.

[0075] Specifically, if there is no clutter, it is assumed that there is air or water inside the target steel structure column; if there is clutter, it is assumed that there is any one of the following impurities inside the target steel structure column: ice, sand, ice-water mixture, or water-sand mixture.

[0076] Step 130: If it exists, then collect the second echo signal inside the target steel structure column; wherein, the second echo signal is a low frequency signal.

[0077] When clutter is present in the first echo signal, traditional high-frequency single probes cannot accurately detect the internal conditions. Furthermore, high-frequency signals have weak penetrating power and experience severe attenuation when penetrating sand or ice, affecting the detection results. Low-frequency signals, on the other hand, attenuate energy more slowly, making them suitable for penetrating thick-walled materials or deep structures. They also have strong anti-interference capabilities, making them ideal for detecting sand or ice inside steel structure columns. Therefore, when acquiring the second echo signal, two independent 50kHz probes from low-frequency module 2 are used. By adjusting the spacing and angle, the sound wave path can be precisely controlled, avoiding obstruction at the probe leading edge, making it suitable for detecting hidden areas inside the target steel structure column.

[0078] Step 140: Based on the first preset gain adjustment method, adjust the gain of the second echo signal to obtain the identification feature map.

[0079] Inside the steel structure, there may be echo overlap issues. This can be addressed by adjusting the gain to separate the overlapping signals. Furthermore, gain adjustment can highlight the signal differences in the second echo interval, making parameters such as the amplitude, phase, frequency, and duration of the second echo signal more apparent.

[0080] After adjusting the gain of the second echo signal, the portion of the waveform in the adjusted second echo signal that clearly indicates the identification characteristics is used as the identification feature map. These identification characteristics may include the amplitude, phase, and frequency of the first wave signal.

[0081] Step 150: Based on the identified feature map, obtain the medium detection results inside the target steel structure column.

[0082] In this embodiment, the waveform characteristics corresponding to different types of media are different. By comparing and identifying the waveform characteristics in the feature map, the type of media inside the target steel structure column can be finally determined.

[0083] In summary, this invention employs a combination of high-frequency and low-frequency signals for judgment, overcoming the limitations of single technologies through multi-dimensional information fusion. This approach balances detection depth and accuracy while enhancing anti-interference capabilities and environmental adaptability. First, based on the first echo signal, an initial measurement of the medium inside the target steel structure column is performed. If clutter is present in the first echo signal, it indicates the possible presence of a mixture, requiring further determination of the type of mixture using the second echo signal. When the gain is too low, weak signals may be ignored; when the gain is too high, noise interference will be excessively amplified. Therefore, by adjusting the gain of the second echo signal, a characteristic echo reflecting the medium's characteristics is obtained. This characteristic echo is then used as a feature map for identification and detection, ultimately determining the type of medium inside the target steel structure column. This method improves detection accuracy and is applicable to the detection of ice, water, and sand inside target steel structure columns. Compared to traditional methods that use different methods for detecting ice and sand, this approach is more cost-effective.

[0084] In an optional embodiment, determining whether clutter exists in the first echo signal in step 120 may include:

[0085] If there is no water accumulation echo in the first echo signal, the gain of the first echo signal is adjusted based on the second preset gain adjustment method.

[0086] Determine whether there is noise in the first echo signal after gain adjustment.

[0087] Figures 3a-3f This is a schematic diagram of the first echo signal corresponding to different media inside the steel structure column provided in an embodiment of the present invention. The following is in conjunction with... Figures 3a-3f This embodiment will be described below. In each figure, the horizontal axis represents the number of time points, and the vertical axis represents the corresponding amplitude. The following is a detailed description of each figure:

[0088] Figure 3a This is a schematic diagram of the first echo signal when the medium inside the steel structure column is air, i.e., when the tube is empty. Figure 3b This is a schematic diagram of the first echo signal when the internal medium of a steel structure column is water. Figure 3c This is a schematic diagram of the first echo signal when the internal medium of a steel structure column is sand and gravel. Figure 3dThis is a schematic diagram of the first echo signal when the internal medium of a steel structure column is a mixture of water and sand. Figure 3e This is a schematic diagram of the first echo signal when the internal medium of a steel structure column is ice. Figure 3f A schematic diagram of the first echo signal corresponding to the water-sand mixture after adjustment using the second preset gain adjustment method.

[0089] When using a high-frequency single probe 10 to probe the interior of a steel structure column, only when there is water accumulation will the first echo signal produce a signal like this. Figure 3b As shown in the water echo diagram, in other cases, when air, sand, or gravel are present, the high-frequency single probe 10 will be unable to detect them. Furthermore, when water and sand are present simultaneously, detecting only the bottom will not reveal the presence of water, easily leading to false positives.

[0090] Therefore, if the first echo signal contains a water accumulation echo, it indicates that the internal medium of the steel structure column is water, and subsequent signal acquisition is unnecessary. If the first echo signal does not contain a water accumulation echo, it indicates that the internal medium of the steel structure column may be one of the following: air, sand, a mixture of water and sand, ice, or a mixture of ice and water. Therefore, further judgment is required.

[0091] If no water accumulation echo is found in the first echo signal, its corresponding current acoustic time is compared with a preset acoustic time threshold. If the error between the current acoustic time and the preset acoustic time threshold is less than the preset error, the internal medium of the steel structure column is considered to be air, and no further steps are required. The preset acoustic time threshold is used to characterize the acoustic time when the internal medium of the steel structure column is air.

[0092] If the error between the current acoustic time and the preset acoustic time threshold is not less than the preset error, the second preset gain adjustment method is used to adjust the gain of the first echo signal without water accumulation echo. If there is no noise in the first echo signal after gain adjustment, it is assumed that the internal medium of the steel structure column is air, and the previous acoustic time judgment was incorrect, so no further signal acquisition is needed. If there is noise in the first echo signal after gain adjustment, it is assumed that the internal medium of the steel structure column may be one of sand, water-sand mixture, ice, or ice-water mixture, so further signal acquisition is needed to determine the specific type of medium.

[0093] In this embodiment, the second preset gain adjustment method can adjust the gain within a preset gain adjustment range by ascending or descending order, or by adjusting the gain according to a preset gain value. During the adjustment process, if noise appears in the first echo signal, the adjustment is stopped, and subsequent signal acquisition is performed. Alternatively, if no noise appears after adjusting the gain within the preset gain adjustment range according to the preset gain value, the adjustment is stopped, and no further signal acquisition is performed.

[0094] In an optional embodiment, step 140 involves adjusting the gain of the second echo signal based on a first preset gain adjustment method to obtain an identification feature map, including:

[0095] Within a preset gain range, the second echo signal is gain-adjusted according to a preset sequence to obtain the second echo signal after gain adjustment.

[0096] The identification feature map is obtained based on the second echo signal after gain adjustment.

[0097] Figures 4a-4b This is an ice recognition feature map based on a preset gain range adjustment provided in an embodiment of the present invention. The following is in conjunction with... Figures 4a-4b This embodiment will now be described.

[0098] This embodiment uses an ice-water mixture as an example to illustrate the process of determining the identification feature map. The identification feature map for other media can be determined based on this embodiment. In the figure, point Y represents the amplitude, and point X represents the first wave time.

[0099] Figure 4a This refers to the second echo signal of ice at a gain of 64dB when adjusted in descending order within a preset gain range. Figure 4b This is the second echo signal of ice at a gain of 47dB. In the graphs, the horizontal axis represents the number of time points, and the vertical axis represents the corresponding amplitude.

[0100] Different media types correspond to different first waves. In this embodiment, the first wave is identified to distinguish different media types. The identification feature map is also the first wave feature map of the second echo signal.

[0101] based on Figure 4a and Figure 4b It can be seen that by continuously adjusting the gain, the characteristics of the first wave waveform can be clearly distinguished at various gain values, for example... Figure 4b The first wave waveform characteristics, compared to Figure 4a The first wave waveform is more obvious.

[0102] Gain adjustment can be done automatically by the equipment or manually.

[0103] If the equipment is automatically adjusted, the signal-to-noise ratio and amplitude stability of the first waveform can be used as adjustment indicators; among them, amplitude stability is used to characterize the fluctuation range of the second echo signal amplitude.

[0104] During the gain adjustment process, the adjustment is performed in a preset order, such as ascending or descending, with each adjustment based on a preset step size, such as 2dB, to obtain the second echo signal after each gain adjustment and to determine the corresponding signal-to-noise ratio and amplitude stability.

[0105] The signal-to-noise ratio (SNR) is compared with a preset SNR threshold, and the fluctuation range of the second echo signal amplitude is compared with a preset fluctuation range. When the SNR is greater than the preset SNR threshold or the fluctuation range of the signal amplitude is within the preset fluctuation range, the adjustment is stopped, and the obtained first waveform is used as the identification feature map.

[0106] The manual adjustment method can be carried out in a preset order, such as ascending or descending. Each adjustment is based on a preset step size. During the adjustment process, the manual judges whether it meets the requirements in order to obtain the recognition feature map.

[0107] In this embodiment, the preset gain range can be obtained based on the gain range of each medium. Specifically, through experiments, the gain range corresponding to the recognition feature map of each medium is determined, and these ranges are comprehensively considered to determine the final preset gain range.

[0108] In an optional embodiment, step 140, which involves adjusting the gain of the second echo signal based on a first preset gain adjustment method to obtain an identification feature map, further includes:

[0109] Based on a preset gain value, the gain of the second echo signal is adjusted to obtain the second echo signal after gain adjustment.

[0110] The identification feature map is obtained based on the second echo signal after gain adjustment.

[0111] In this embodiment, the gain of the second echo signal can also be adjusted based on a preset gain value. During the adjustment process, both automatic adjustment by the device and manual adjustment can be selected. During the adjustment process, the signal-to-noise ratio and amplitude stability of the first waveform can also be used as adjustment indicators. For details, please refer to the above-mentioned related embodiments.

[0112] In this embodiment, the preset gain threshold can be determined based on the gain threshold corresponding to the recognition feature map of different media. For example, the recognition feature map of ice performs best at a gain of 76dB, the recognition feature map of sand and gravel performs best at a gain of 46dB, the recognition feature map of ice-water mixture performs best at a gain of 53dB, and the recognition feature map of water-sand mixture performs best at a gain of 81dB. Therefore, the preset gain threshold can be selected from 46dB, 53dB, 76dB, and 81dB.

[0113] Based on a preset gain threshold, the gain of the second echo signal is adjusted to obtain the first waveform of the second echo signal under different preset gain values. The first waveform under different preset gain values ​​is judged by indicators, and the first waveform that best meets the indicators is selected as the identification feature map.

[0114] It should be understood that the specific gain values ​​given above are merely illustrative examples and do not constitute any limitation on the preset gain threshold. In practical applications, due to the different thicknesses of ice voids and sand and gravel, a medium may correspond to more than one gain value. Therefore, the preset gain threshold is not necessarily limited to only the four mentioned above.

[0115] In an optional embodiment, step 150, based on the identified feature map, obtains the medium detection result inside the target steel structure column, including:

[0116] The recognition feature map is input into the pre-trained first target detection model to obtain the medium detection results inside the target steel structure column.

[0117] The test results for the medium inside the target steel structure column include one of the following: water, air, sand, ice, ice-water mixture, and water-sand mixture.

[0118] Figures 5a-5f These are identification feature maps of different media provided in the embodiments of the present invention; the following is in conjunction with... Figures 5a-5f This embodiment will now be described.

[0119] Figure 5a This is a feature map for identifying water accumulation; Figure 5b This is a feature map for identifying ice. Figure 5c This is a feature map for identifying sand and gravel. Figure 5d A feature map for identifying water-sand mixtures; Figure 5e This is a feature map for identifying air. Figure 5f This is a feature map for identifying ice-water mixtures. In each graph, the horizontal axis represents the number of time points, and the vertical axis represents the corresponding amplitude.

[0120] based on Figures 5a-5f It is known that different media have different identification feature maps. The amplitude, frequency and waveform in the identification feature map can be used as identification features. The pre-trained first target detection model is used to identify each identification feature map in order to obtain the medium detection results inside the target steel structure column.

[0121] In this embodiment, the pre-trained first object detection model can be a YOLO series model or a neural network model. The pre-trained first object detection model is trained by using the recognition feature maps of each medium as input and the type of each medium as output.

[0122] In an optional embodiment, if the medium detection result inside the target steel structure column is ice or an ice-water mixture, the method further includes:

[0123] The recognition feature map is input into a pre-trained second target detection model to determine the thickness of the ice detachment.

[0124] Figures 6a-6f These are identification feature diagrams of ice and ice-water mixtures with different delamination thicknesses provided in embodiments of the present invention; the following is in conjunction with... Figures 6a-6f This embodiment will be described.

[0125] Figures 6a-6f It is a recognition feature map obtained based on a preset gain value, which is 76dB. Figure 6a This is a feature map of ice under 8mm delamination. Figure 6b This is a feature map of ice under 15mm delamination. Figure 6c This is a feature map of ice under 20mm delamination. Figure 6d This is a feature map of ice under 30mm delamination. Figure 6e This is a feature map of ice under 40mm delamination. Figure 6f This is a feature diagram of the ice-water mixture under 40mm de-cavitation.

[0126] Based on the above-mentioned ice identification feature maps with different delamination thicknesses and Figure 5b As shown in the identification feature map of ice without voids, when the void thickness is small (less than 30 mm), the detected waveform is closer to the waveform of ice in its intact state. When the void thickness increases (greater than 30 mm), the difference in the detected waveform becomes greater. Therefore, when the void is small, the void thickness of ice cannot be determined manually from the identification feature map; when the void is large (e.g., 40 mm), the identification feature map of the ice-water mixture and ice is not significantly different.

[0127] Based on the above analysis, the detection method provided in this embodiment of the invention has discovered the following patterns in the detection of ice:

[0128] (1) When the ice is intact or the void is small, its first wave signal has a strong recognition effect and can accurately identify the internal ice formation.

[0129] (2) With a fixed gain, as the ice detachment increases, a concentrated characteristic wave gradually appears at the front end of the waveform, which is most obvious when the ice detachment is 40mm.

[0130] (3) The ice-water mixture has no effect on the test results.

[0131] Since it is difficult to accurately identify waveform differences in the feature map when the void is small, machine identification can be used to determine the void thickness corresponding to the ice's feature map. Specifically, the amplitude, frequency, and waveform in the ice's feature map can be used as identification features. A pre-trained second target detection model can then be used to identify each feature map to obtain the medium detection results inside the target steel structure column.

[0132] In this embodiment, the pre-trained second object detection model is the same as the pre-trained first object detection model, which can be a YOLO series model or a neural network model. Unlike the pre-trained first object detection model, the pre-trained second object detection model is trained by using the recognition feature maps of ice under different voids as input and the void thickness of the ice as output.

[0133] Taking the YOLO series models as an example, the following explains how to train a second object detection model.

[0134] Points were marked on the outer surface of the steel structure column, and second echo signals with different void thicknesses were collected at different points as echo signals inside the steel structure column to obtain the identification feature map of each second echo signal, and these were used as a dataset.

[0135] Use annotation tools (such as LabelImg) to annotate the void thickness corresponding to each recognition feature map in the dataset for training and testing of void thickness recognition.

[0136] The dataset is divided into a dataset, a test set, and a validation set according to a certain ratio.

[0137] Using the YOLOv5 training script, specify parameters such as the configuration file and pre-training weights, and begin training the model. During training, tools such as TensorBoard can be used to monitor metrics such as changes in the loss function and the accuracy of the validation set to evaluate the model's training performance.

[0138] The suitability of a model for identification is determined by detection precision and average detection precision (mAP). Detection precision metrics include accuracy (P) and recall (R), while model execution speed is evaluated using detection precision and mAP.

[0139] The formula for calculating accuracy is as follows:

[0140]

[0141] In the formula: The number of samples for which the model successfully identifies image features and detects the correct internal state. This represents the number of samples for which the model failed to detect the internal state.

[0142] The formula for calculating recall rate is as follows:

[0143]

[0144] In the formula, The number of samples for which the model incorrectly detected internal states.

[0145] The formula for calculating the average accuracy mAP is as follows:

[0146]

[0147] In the formula, It is the total number of categories, AP i It is the first i AP value for the category.

[0148] The above describes the training process for the second object detection model. The training process for the first object detection model can be referenced from this process.

[0149] In an optional embodiment, the method further includes:

[0150] Collect temperature data on the surface of the target steel structure column.

[0151] The time-frequency characteristic data are determined based on the second echo signal after gain adjustment.

[0152] The time-frequency characteristics of the target are determined based on temperature data.

[0153] The degree of waveform distortion is determined based on the target time-frequency characteristic data and the time-frequency characteristic data.

[0154] The detection results of the medium inside the target steel structure column are determined based on the degree of waveform distortion.

[0155] In this embodiment, the stress generated by ice expansion alters the ultrasonic wave propagation path, causing waveform distortion. Since the distortion degree of ice is much greater than that of other media, this embodiment pre-constructs a corresponding correlation model based on the relationship between time-frequency characteristic data and temperature data. Based on the correlation model, the target time-frequency characteristic data corresponding to the temperature data is determined to ascertain the degree of waveform distortion, thereby determining whether the medium inside the target steel structure column is ice or another medium.

[0156] The method provided in this embodiment differs from the other related embodiments described above. The method provided in this embodiment can only detect whether the medium inside the target steel structure column is ice. This method is applicable to steel structure facilities in low-temperature environments or at risk of freezing, to investigate internal damage to the steel structure caused by freezing, and to monitor whether there is a risk of frost heave inside the steel structure in winter.

[0157] In an optional embodiment, if the medium detection result inside the target steel structure column is sand or a mixture of sand and water, the method further includes:

[0158] Obtain the propagation speed of the second echo signal within the target steel structure column.

[0159] The target calculation formula is determined based on the propagation speed.

[0160] The thickness of the sand and gravel is determined based on the propagation speed and target calculation formula.

[0161] Figures 7a-7f These are identification feature maps of sand and gravel of different thicknesses provided in embodiments of the present invention; wherein, Figure 7a Identification feature diagram of sand and gravel with a thickness of 5mm; Figure 7b Identification feature map of sand and gravel with a thickness of 10mm; Figure 7c Identification feature map of sand and gravel with a thickness of 20mm; Figure 7d Identification feature map of sand and gravel with a thickness of 25mm; Figure 7e Identification feature map of sand and gravel with a thickness of 30mm; Figure 7f This is a feature map for identifying sand and gravel with a thickness of 40mm. Point Y in the map represents the amplitude, and point X represents the first wave time. The frequency used is 100MHz, and one point is taken every 0.01µs. Figure 7b With the X-coordinate at 9948, the acoustic time is 99.48 μs. The corresponding propagation speed of the second echo signal within the target steel structure column is:

[0162]

[0163] In the formula, The propagation speed of the second echo signal within the target steel structure column; The inner diameter of the target steel structure column; When it is a sound; The thickness of the target steel structure column; Let be the propagation speed of the second echo signal in the steel, and be a known quantity.

[0164] Given the acoustic time, inner diameter, thickness, and the propagation speed of the second echo signal in the steel, the propagation speed of the second echo signal in the target steel structure column can be calculated using the above formula.

[0165] Based on the experiment, the propagation speed of ultrasound in sand and gravel of different thicknesses is shown in Table 1:

[0166] Table 1. Propagation speed of ultrasound in sand and gravel of different thicknesses

[0167]

[0168] Based on Table 1, the relationship between sand thickness and the propagation speed of the second echo signal within the target steel structure column can be obtained as follows:

[0169] First calculation formula:

[0170]

[0171] Second calculation formula:

[0172]

[0173] In the formula, The thickness of the sand and gravel.

[0174] That is, the corresponding calculation formula is selected based on the propagation speed of the second echo signal within the target steel structure column and a preset speed threshold, which is obtained experimentally. When the propagation speed of the second echo signal within the target steel structure column is less than or equal to the first preset speed threshold, the first calculation formula is used as the target calculation formula; when the propagation speed of the second echo signal within the target steel structure column is greater than the first preset speed threshold and less than or equal to the second preset speed threshold, the second calculation formula is used as the target calculation formula.

[0175] The first preset speed threshold is less than the second preset speed threshold.

[0176] By substituting the propagation speed of the second echo signal within the target steel structure column into the target calculation formula, the thickness of the sand and gravel can be determined.

[0177] Alternatively, in an optional embodiment, if the medium detection result inside the target steel structure column is sand or a mixture of sand and water, the method further includes:

[0178] Obtain the propagation speed of the second echo signal within the target steel structure column.

[0179] The initial thickness of the sand and gravel is determined based on the propagation speed and the formula for calculating the first target.

[0180] Determine whether the initial thickness of the sand and gravel meets the constraints;

[0181] If the conditions are not met, the thickness of the sand and gravel is determined based on the propagation speed and the formula for calculating the second target.

[0182] Since selecting the appropriate calculation formula and calculating the thickness of sand and gravel based on the propagation speed of the second echo signal in the target steel structure column may increase the error, this embodiment can also judge whether the calculation result is accurate based on the initial thickness of the sand and gravel.

[0183] The formula for calculating the first objective is as follows:

[0184]

[0185] The formula for calculating the second objective is:

[0186]

[0187] The constraints are the thickness range of sand and gravel corresponding to the calculation formulas for the first and second objectives.

[0188] If the initial thickness of the sand and gravel determined according to the propagation speed and the first target calculation formula satisfies the constraint condition that it is less than or equal to 5, then the initial thickness of the sand and gravel shall be taken as the thickness of the sand and gravel.

[0189] In an optional embodiment, if the medium detection result inside the target steel structure column is a water-sand mixture, the method further includes:

[0190] Time-frequency analysis was performed on the second echo signal after gain adjustment to identify the coherent peaks in the second echo signal.

[0191] Determine the characteristic parameters corresponding to the coherent peak, including frequency, duration, and amplitude.

[0192] Based on the characteristic parameters, the propagation speed of the second echo signal in the water-sand mixture and the thickness of the sand are determined.

[0193] In the actual service environment of steel structure columns, static media accumulation is not the only issue. For example, rainwater can carry sand and gravel particles into the column, creating a dynamic flow of water and sand mixture. In this situation, some of the sand and gravel settles at the bottom of the steel structure column, while the rest mixes with the accumulated water and remains on top of the settled sand and gravel. The aforementioned embodiments are used to measure the thickness of static media, but they are ineffective in handling this multi-media dynamic fluid-structure interaction scenario, failing to obtain key parameters such as sand and gravel particle concentration and flow velocity, making it difficult to comprehensively assess the risk of erosion from the media inside the column. Therefore, this embodiment provides another method for detecting the flow velocity and concentration of sand and gravel under dynamic conditions.

[0194] Specifically, when ultrasonic waves, or second echo signals, are incident on a flowing mixture of water and sand, individual sand particles act as scatterers, causing the ultrasonic waves to be scattered. Because the particles are in dynamic flow, the frequency and phase of the scattered waves change in real time with the particle's position. Scattered waves from particles at different positions and in different states of motion superimpose to form a complex scattered wave field.

[0195] During fluid flow, sand and gravel particles may agglomerate locally due to Brownian motion and fluid shear forces. The agglomerated particle group is equivalent to a "large-sized scatterer," and its scattering characteristics differ significantly from those of a single particle. At this point, the scattered waves form a brief coherent peak due to the cooperative scattering of the particle group. The frequency of the coherent peak is related to the equivalent size of the particle agglomeration, while the duration is related to the duration of the agglomeration and the flow velocity.

[0196] Therefore, this embodiment performs time-frequency analysis on the second echo signal, identifies the coherent peaks in the second echo signal, and determines the characteristic parameters corresponding to the coherent peaks, such as frequency, duration, and amplitude.

[0197] Using computational fluid dynamics software, the flow state of water-sand mixtures inside steel structure columns was simulated, and flow field data under different flow velocities, particle concentrations, and agglomeration characteristics were obtained. Simultaneously, combined with acoustic finite element simulation, flow field parameters such as particle position, velocity, and agglomeration size were used as boundary conditions to simulate the propagation process of ultrasonic waves in this complex fluid-structure interaction environment. Features such as the coherence peak frequency and duration of the scattered waves were extracted, and a database mapping the characteristic parameters to sand and gravel parameters was established.

[0198] The obtained characteristic parameters are compared with a pre-established database of characteristic parameters and sand parameters to determine the propagation speed of the second echo signal in the water-sediment mixture and the thickness of the sand and gravel deposited at the bottom. Then, the concentration of the water-sediment mixture mixed with the accumulated water is determined based on the propagation speed of the second echo signal in the water-sediment mixture.

[0199] Therefore, this embodiment can monitor the dynamic media intrusion caused by rainwater erosion in real time. By controlling the concentration and flow rate of sand and gravel particles, it can provide early warnings of problems such as wear and accelerated corrosion of the column inner wall caused by media flow, guide targeted protection and maintenance measures, extend the service life of steel structure columns, and ensure structural safety.

[0200] In summary, this invention employs a combination of high-frequency and low-frequency signals for judgment, overcoming the limitations of single technologies through multi-dimensional information fusion. This approach balances detection depth and accuracy while enhancing anti-interference capabilities and environmental adaptability. First, based on the first echo signal, an initial measurement of the medium inside the target steel structure column is performed. If clutter is present in the first echo signal, it indicates the possible presence of a mixture, requiring further determination of the type of mixture using the second echo signal. When the gain is too low, weak signals may be ignored; when the gain is too high, noise interference will be excessively amplified. Therefore, by adjusting the gain of the second echo signal, a characteristic echo reflecting the medium's characteristics is obtained. This characteristic echo is then used as a feature map for identification and detection, ultimately determining the type of medium inside the target steel structure column. This method improves detection accuracy and is applicable to the detection of ice, water, and sand inside target steel structure columns, saving detection costs.

[0201] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0202] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0203] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for detecting the internal medium of a steel structure column, characterized in that, include: The first echo signal inside the target steel structure column is acquired; wherein, the first echo signal is a high-frequency signal; Determine whether there is noise in the first echo signal; If present, the second echo signal inside the target steel structure column is acquired; wherein the second echo signal is a low-frequency signal. Based on the first preset gain adjustment method, the gain of the second echo signal is adjusted to obtain the identification feature map; Based on the identified feature map, the medium detection result inside the target steel structure column is obtained; The step of adjusting the gain of the second echo signal based on a first preset gain adjustment method to obtain a recognition feature map includes: Within a preset gain range, the second echo signal is gain-adjusted according to a preset sequence to obtain a gain-adjusted second echo signal; wherein, during the gain adjustment process, the signal-to-noise ratio and amplitude stability of the first waveform are used as adjustment indicators; the amplitude stability is used to characterize the fluctuation range of the amplitude of the second echo signal; The signal-to-noise ratio (SNR) is compared with a preset SNR threshold, and the fluctuation range of the second echo signal amplitude is compared with a preset fluctuation range. When the SNR is greater than the preset SNR threshold or the fluctuation range of the signal amplitude is within the preset fluctuation range, the adjustment is stopped, and the first waveform of the second echo signal after gain adjustment is used as the identification feature map. If the test result of the medium inside the target steel structure column is ice or an ice-water mixture, the method further includes: The identified feature map is input into a pre-trained second target detection model to determine the thickness of the ice detachment.

2. The method for detecting the internal medium of a steel structure column according to claim 1, characterized in that, The determination of whether there is noise in the first echo signal includes: If there is no water accumulation echo in the first echo signal, the gain of the first echo signal is adjusted based on the second preset gain adjustment method. Determine whether there is noise in the first echo signal after gain adjustment.

3. The method for detecting the internal medium of a steel structure column according to claim 1, characterized in that, Alternatively, the step of adjusting the gain of the second echo signal based on a first preset gain adjustment method to obtain a recognition feature map includes: Based on a preset gain value, the gain of the second echo signal is adjusted to obtain the second echo signal after gain adjustment. Based on the second echo signal after gain adjustment, a recognition feature map is obtained.

4. The method for detecting the internal medium of a steel structure column according to claim 3, characterized in that, The step of obtaining the medium detection result inside the target steel structure column based on the identified feature map includes: The identification feature map is input into a pre-trained first target detection model to obtain the medium detection result inside the target steel structure column; The medium detection results inside the target steel structure column include one of the following: water, air, sand, ice, ice-water mixture, and water-sand mixture.

5. The method for detecting the internal medium of a steel structure column according to claim 4, characterized in that, If the test result of the medium inside the target steel structure column is sand or gravel or a mixture of water and sand, the method further includes: Obtain the propagation speed of the second echo signal within the target steel structure column; The target calculation formula is determined based on the propagation speed; The thickness of the sand and gravel is determined based on the propagation speed and the target calculation formula.

6. The method for detecting the internal medium of a steel structure column according to claim 4, characterized in that, If the test result of the medium inside the target steel structure column is a mixture of water and sand, the method further includes: Time-frequency analysis was performed on the second echo signal after gain adjustment to identify the coherent peaks in the second echo signal; Determine the characteristic parameters corresponding to the coherent peak, wherein the characteristic parameters include frequency, duration, and amplitude; Based on the characteristic parameters, the propagation speed of the second echo signal in the water-sand mixture and the thickness of the sand are determined.

7. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 6.

8. A system for detecting the internal medium of a steel structure column, characterized in that, It includes a high-frequency module, a low-frequency module, and the electronic device as described in claim 7; wherein the high-frequency module includes a high-frequency single probe, and the low-frequency module includes a low-frequency dual probe.

Images