A method and system for detecting cavities behind subway shield tunnel walls

Abstract

The present invention relates to a method and system for detecting voids behind subway shield tunnel walls, belonging to the technical field of void detection. The method comprises: marking points to be detected on subway tunnel segments with a marker pen; striking the points with a rebound hammer and collecting vibration signals from the points using an accelerometer; denoising the vibration signals at each point to obtain a denoised vibration signal; performing time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal; and completing the detection of voids behind subway shield tunnel walls based on the time domain signal of the denoised vibration signal and the frequency domain signal. By denoising and performing frequency domain analysis on the collected data, the present invention can effectively improve the accuracy of void detection results and reduce the possibility of misjudgment.

Description

Technical Field

[0001] The present invention relates to the technical field of cavity detection, and in particular to a method and system for detecting cavities behind a subway shield tunnel wall. Background Art

[0002] During tunnel service, small cavities gradually form behind the tunnel lining due to factors such as stratum lithology, geological structure, and groundwater. Furthermore, the surrounding rock or backfill behind the lining may contain easily soluble or corrosive substances. Over time, these small cavities connect and expand, eventually evolving into larger cavities. Furthermore, tunnel construction is subject to complex and variable geological conditions, resulting in a harsh construction environment. Failure to ensure standardized and rigorous construction procedures, such as over-excavation, inadequate backfill, and support voids, can lead to voids and cavitation defects in tunnels of varying sizes and locations after composite lining construction. The presence of cavities significantly impacts the ultimate bearing capacity and failure modes of the lining structure. Under the influence of factors such as temperature fluctuations, groundwater hydrostatic pressure, and falling block loads in the cavity area, cavities can lead to reduced bearing capacity, cracks, and spalling of the lining, significantly impacting the safety and durability of the tunnel and shortening its service life. As a primary mode of urban transportation, the structural health of the subway is paramount. Most subways are built within underground tunnels. Cavities exist in the soil behind the lining walls, making them invisible and concealed, placing higher demands on detection methods. Over the past few decades, a series of nondestructive testing methods, such as ground-penetrating radar and infrared thermal imaging, have been researched and applied to subway tunnel defect detection.

[0003] 1) Ground Penetrating Radar

[0004] The detection effectiveness of ground-penetrating radar (GPR) is related to electrical parameters such as the pulse width of the transmitting radar and the dielectric constant, conductivity, and magnetic permeability of the medium. Currently, GPR is widely used to detect voids behind tunnel linings, initial lining thickness, lining leakage, and hidden cracks behind or within the lining. However, the steel bars embedded in the tunnel lining act as strong scatterers, generating clutter in the GPR profile, interfering with the target signal and making the echo of hidden defects difficult to distinguish. Typically, the spacing between steel bars in a standard tunnel lining is approximately 30-40 cm. In contrast, precast shield tunnel sections are embedded with a denser double-layer steel mesh, with a spacing of only approximately 10-15 cm. The image is easily affected by interference from the lining material, resulting in misjudgment.

[0005] 2) Infrared thermal imaging

[0006] As a nondestructive testing method, infrared thermal imaging has gained widespread attention in the field of concrete structure testing due to its advantages such as large testing area, non-contact operation, and simple operation. It has also been gradually applied to on-site inspections of tunnels and bridges. Infrared thermal imaging is often used to inspect reinforced concrete objects. When the depth of the defect exceeds 10mm, heating conditions are required. In addition, the time it takes to inspect the object after heating will also affect the formation of the sample image, which in turn affects the test results. Infrared thermal imaging images cannot accurately reflect the shape and size of the defect, and detection accuracy still needs to be improved. Subway tunnels are located in an underground environment with small temperature fluctuations. The pipe segments are made of reinforced concrete and are easily disturbed. Summary of the Invention

[0007] To solve the above problems, an object of the embodiments of the present invention is to provide a method and system for detecting cavities behind the wall of a subway shield tunnel.

[0008] A method for detecting cavities behind a subway shield tunnel wall, comprising:

[0009] Step 1: Use a marker to mark the points to be inspected on the subway tunnel segments;

[0010] Step 2: Hit the point with a rebound hammer and use an accelerometer to collect the vibration signal of the point;

[0011] Step 3: De-noise the vibration signal at each point to obtain a denoised vibration signal;

[0012] Step 4: Perform time domain conversion on the denoised vibration signal to obtain the frequency domain signal of the corresponding vibration signal;

[0013] Step 5: Complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

[0014] Preferably, the step 3 of performing denoising on the vibration signal of each point to obtain a denoised vibration signal includes:

[0015] Step 3.1: Set the decomposition scale and decompose the vibration signal of each point according to the set decomposition scale to obtain the wavelet coefficients;

[0016] Step 3.2: Determine the denoising threshold based on the set decomposition scale;

[0017] Step 3.3: When the wavelet coefficient is greater than or equal to the denoising threshold, the wavelet coefficient is processed using a processing function to generate processed wavelet coefficients;

[0018] Step 3.4: When the wavelet coefficient is less than the denoising threshold, the corresponding wavelet coefficient is changed to 0;

[0019] Step 3.5: reconstructing the processed wavelet coefficients to form a denoised vibration signal;

[0020] Step 3.6: Use fusion indicators to evaluate the denoised vibration signal;

[0021] Step 3.7: When the weighted evaluation value is not within the preset range, reset the decomposition scale until the weighted evaluation value is within the preset range.

[0022] Preferably, the step 3.3: when the wavelet coefficient is greater than or equal to the denoising threshold, processing the wavelet coefficient using a processing function to generate processed wavelet coefficients includes:

[0023] Using the formula:

[0024]

[0025] The wavelet coefficients are processed to generate the processed wavelet coefficients; where w i ' ,k Represents the processed wavelet coefficients, sign represents the sign function, w i,k represents the kth original wavelet coefficient at the i-th decomposition scale, λ represents the denoising threshold, N represents the length of the vibration signal, and σ represents the noise standard deviation.

[0026] Preferably, the step 3.6: evaluating the denoised vibration signal using a fusion index includes:

[0027] Step 3.6.1: Perform a preliminary evaluation of the denoised vibration signal to obtain the signal's root mean square error (RMS) and signal-to-noise ratio (SNR).

[0028] Step 3.6.2: Normalize the root mean square error index and the signal-to-noise ratio index to obtain the normalized root mean square error index and signal-to-noise ratio index;

[0029] Step 3.6.3: Calculate the weights of the normalized root mean square error and signal-to-noise ratio indicators;

[0030] Step 3.6.4: Use the weights to weight the root mean square error index and the signal-to-noise ratio index to obtain the fusion index.

[0031] Preferably, in step 3.6.1, the calculation formulas for the root mean square error index and the signal-to-noise ratio index are:

[0032]

[0033] Where x(i) represents the original vibration signal, y(i) represents the signal-to-noise ratio, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the root mean square error.

[0034] Preferably, the step 3.6.3: calculating the weights of the normalized root mean square error index and the signal-to-noise ratio index includes:

[0035] Calculate the ratio of each indicator to all indicators and determine the weight based on the ratio; the weight calculation formula is:

[0036]

[0037] Among them, P ij It represents the ratio of the jth index when the decomposition scale is i, x ij It represents the value of the jth indicator when the decomposition layer is i after normalization, m represents the number of decomposition scales, and w j Represents the weight of the j-th indicator.

[0038] Preferably, the step 4: performing time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal includes:

[0039] Fast Fourier transform is used to transform the denoised vibration signal to obtain the frequency domain signal of the corresponding vibration signal; the conversion formula is:

[0040]

[0041] Where c≥0, e v (c) = x(2c), o d (c) = x(2c+1), N is the total number of sampling points, k is the frequency index used to represent different frequency components, X k1 Represents the first N / 2 frequency components of the final Fourier transform result, Represents the last N / 2 frequency components of the Fourier transform result, k1 represents the frequency index, which determines the position of the frequency component in the frequency domain, e v (c) represents the even part, that is, the sampling points of all even index positions in signal c, x(2c) represents extracting the sample values ​​of the signal and rearranging them so that the indexes of these sample values ​​are even indexes, o d (c) represents the odd part, that is, the sampling points of all odd index positions in signal c, DFT represents the discrete Fourier transform function, represents the rotation factor.

[0042] Preferably, the step 5: completing the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal, includes:

[0043] The amplitude of the periodic fluctuation phase in the time domain signal is extracted. When the amplitude is greater than the set threshold, a cavity appears at the corresponding point in the subway tunnel.

[0044] The peak value in the frequency domain signal is extracted. When the peak value exceeds the preset threshold, a hole appears at the corresponding point in the subway tunnel.

[0045] The present invention also provides a system for detecting cavities behind subway shield tunnel walls, comprising:

[0046] Point marking module, used to mark the points to be inspected on the tube segments of subway tunnels with a marker pen;

[0047] A vibration signal acquisition module is used to strike a point and collect the vibration signal of the point using an acceleration sensor;

[0048] A denoising module is used to denoise the vibration signal of each point to obtain a denoised vibration signal;

[0049] A time domain conversion module is used to perform time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal;

[0050] The cavity detection module is used to complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

[0051] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for detecting cavities behind the wall of a subway shield tunnel are implemented.

[0052] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0053] The present invention relates to a method for detecting cavities behind the wall of a subway shield tunnel. Compared with the prior art, the present invention can effectively improve the accuracy of cavity detection results and reduce the possibility of misjudgment by performing denoising processing and frequency domain analysis on collected data.

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0056] Figure 1 A flow chart of a method for detecting cavities behind a subway shield tunnel wall provided by the present invention;

[0057] Figure 2 This is a simulation effect diagram of continuous voids provided by the present invention;

[0058] Figure 3 Schematic diagram of the lining segment area division provided by the present invention;

[0059] Figure 4 This is a schematic diagram of a time domain signal without holes in area ② provided by the present invention;

[0060] Figure 5 A schematic diagram of the time domain signals of the corresponding points where holes are placed in area ② provided by the present invention;

[0061] Figure 6 This is a schematic diagram of a time domain signal without holes in area ③ provided by the present invention;

[0062] Figure 7 Schematic diagram of the time domain signal with a hole placed in area ③ provided by the present invention;

[0063] Figure 8 Schematic diagram of frequency signals without holes in area ② provided by the present invention;

[0064] Figure 9 Schematic diagram of frequency signals of points corresponding to the placement of holes in area ② provided by the present invention;

[0065] Figure 10 This is a schematic diagram of frequency signals in area ③ provided by the present invention where no holes are placed;

[0066] Figure 11 Schematic diagram of frequency signals with holes placed in area ③ provided by the present invention. DETAILED DESCRIPTION

[0067] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0068] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0069] In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediary; or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0070] To address the problem that existing detection methods are susceptible to the influence of concrete and steel bars, this paper proposes a method for detecting voids in rock behind a wall based on the impact echo method. An accelerometer is mounted on a test surface near a pulse source to receive vibration data, generating a voltage-time signal (i.e., a waveform) that is recorded by a data acquisition device. This signal describes transient vibrations caused by multiple reflections of stress waves within the structure at interfaces such as concrete / air and concrete / rebar. The dominant frequencies in these vibrations are related to stress waves reflected from different depths within the structure. Fast Fourier transform (FFT) signal processing is used to obtain frequency domain data. Time domain data is combined with frequency domain data to better capture the characteristics of voids. A comprehensive analysis of the influence of cavity depth, cavity size, relative position of the cavity to lining bolt holes and grouting holes, and boundary conditions is conducted, clarifying the key detection points and applicable cavity types based on the impact echo method. This method is of great significance for the application of the impact echo method in the detection of defects in rock behind a wall and for improving the accuracy and efficiency of detecting voids in rock behind a wall.

[0071] See also Figure 1, a method for detecting voids behind a subway shield tunnel wall, comprising:

[0072] Test site layout: Use a crane to place the segments side by side on the ground, and build a T-shaped wall to ensure the stability of the segments;

[0073] One month after the segments were in place, soil was filled behind them. To ensure soil density, each layer was compacted with a roller before the next layer was laid. When the filled soil reached the required cavity height, empty bottles were placed to simulate the cavity. Using a shovel, soil was carefully spread onto and around the empty bottles, compacting the soil around them. During this time, care was taken to ensure the bottles remained stationary. Finally, the entire area behind the segments was paved.

[0074] The segments used in this invention are standard segments for subway shield tunnels, with parameters provided by the manufacturer: 3.4m long, 1.2m wide, and 300mm thick. The actual segments are curved, with an inner diameter of 5.4m and an outer diameter of 6.1m. The concrete strength is C50, and the waterproof grade is P10.

[0075] In order to minimize the influence of the material of the object used to simulate the void on the experiment, the present invention uses water bottles of different capacities to simulate voids of different sizes, as shown in Table 1. In order to simulate the existence of continuous voids, according to Figure 2 As shown, two 1.5-liter mineral water bottles are stacked up and down, and two 5-liter water bottles are placed side by side.

[0076] Table 1 Dimensions of water bottles used to simulate voids

[0077] volume Bottom diameter (cm) Water bottle height (cm) 1.5L 8 16 5L 15 33 12L 23 36 Continuous voids 15-16 82

[0078] The detailed experimental settings are shown in Table 2. This experiment uses the cavity depth and cavity size as variables, and sets up experiments 1-3 respectively. Considering that there are bolt holes and grouting holes on the surface of the segment, and because the segment is placed vertically on the ground, the vibration signals between points at different heights on the segment may be different. Therefore, this paper adds a set of variables "placement area (R)" as experiment 1, as follows Figure 3 As shown in the figure, the segment is divided into two regions, ① and ②, based on the location of the grouting holes on the segment surface. When the cavity is placed behind R-②, it is placed behind the grouting holes. Regions ① and ③ correspond to two different heights of the segment, with region ③ being closer to the ground.

[0079] Table 2 Experimental settings

[0080]

[0081] Note: *Burial depth: The distance between the cavity and the segment. 0 cm means the cavity is directly adjacent to the segment wall, while 16.5 cm means the cavity is 16.5 cm away from the segment wall. **During pre-buried cavities, the original intention in Experiment 3 was to maintain a 15 cm distance between the cavity and the segment, but this was not possible in practice, resulting in a final measured distance of 16.5 cm.

[0082] In Experiment 1, holes with the same properties were placed behind R-①, ②, and ③: air hole, burial depth 0, and hole size 5 L, corresponding to B-1, B-3, and C-1 in Table 3. The purpose of Experiment 2 was to obtain hole signals of different sizes at the same location. The four holes were all placed close to the segment wall (ensuring the same burial depth). The hole sizes were 1.5 L, 5 L, 12 L, and continuous, respectively, and the corresponding holes were numbered D-1, B-3, E-1, and E-2. The purpose of Experiment 3 was to obtain hole signals at different burial depths. The distances between the holes and the segment wall were set to 0 cm, 16.5 cm, and 30 cm, and the corresponding holes were numbered B-3, D-2, and C-2. No holes were placed in the control group (segment A).

[0083] Table 3 Cavity properties behind different segments

[0084]

[0085] Step 1: Use a marker to mark the points to be inspected on the subway tunnel segments;

[0086] Step 2: Hit the point with a rebound hammer and use an accelerometer to collect the vibration signal of the point;

[0087] In practical applications, the present invention requires marking the points to be detected on the pipe segment with a marker pen, placing the acceleration sensor close to the point, and hitting the point with a rebound hammer. At this time, the collector collects the signal of the acceleration sensor and transmits it to the computer synchronously. The signal is then stored on the computer. The signal of each point is named: pipe segment number-horizontal coordinate value-vertical coordinate value. After the data collection of the points near the sensor is completed, the sensor is removed and pasted near the remaining points, and then struck and signal collection is carried out. Repeat the above process until the data of all points are collected. Among them, the acceleration sensor model is the single-axis sensor of Langs; the signal collector model is the Yangzhou Jingming dynamic acquisition module JM5938; the rebound hammer uses the high-strength concrete digital display rebound hammer HT450-D of Zhuolin Technology.

[0088] Step 3: De-noise the vibration signal at each point to obtain a denoised vibration signal;

[0089] Furthermore, step 3 includes:

[0090] Step 3.1: Set the decomposition scale and decompose the vibration signal of each point according to the set decomposition scale to obtain the wavelet coefficients;

[0091] Step 3.2: Determine the denoising threshold based on the set decomposition scale;

[0092] Step 3.3: When the wavelet coefficient is greater than or equal to the denoising threshold, the wavelet coefficient is processed using a processing function to generate processed wavelet coefficients;

[0093] In step 3.3, the present invention may adopt the formula:

[0094]

[0095] The wavelet coefficients are processed to generate the processed wavelet coefficients; where w i ' ,k Represents the processed wavelet coefficients, sign represents the sign function, w i,k represents the kth original wavelet coefficient at the i-th decomposition scale, λ represents the denoising threshold, N represents the length of the vibration signal, and σ represents the noise standard deviation.

[0096] Step 3.4: When the wavelet coefficient is less than the denoising threshold, the corresponding wavelet coefficient is changed to 0;

[0097] Step 3.5: reconstructing the processed wavelet coefficients to form a denoised vibration signal;

[0098] Step 3.6: Use fusion indicators to evaluate the denoised vibration signal;

[0099] Furthermore, step 3.6 includes:

[0100] Step 3.6.1: Perform a preliminary evaluation of the denoised vibration signal to obtain the signal's root mean square error (RMS) and signal-to-noise ratio (SNR).

[0101] In step 3.6.1, the calculation formulas for the root mean square error index and the signal-to-noise ratio index are:

[0102]

[0103] Where x(i) represents the original vibration signal, y(i) represents the signal-to-noise ratio, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the root mean square error.

[0104] Step 3.6.2: Normalize the root mean square error index and the signal-to-noise ratio index to obtain the normalized root mean square error index and signal-to-noise ratio index;

[0105] Step 3.6.3: Calculate the weights of the normalized root mean square error and signal-to-noise ratio indicators;

[0106] Among them, step 3.6.3 includes:

[0107] Calculate the ratio of each indicator to all indicators and determine the weight based on the ratio; the weight calculation formula is:

[0108]

[0109] Among them, P ij It represents the ratio of the jth index when the decomposition scale is i, x ij It represents the value of the jth indicator when the decomposition layer is i after normalization, m represents the number of decomposition scales, and w j Represents the weight of the j-th indicator.

[0110] Step 3.6.4: Use the weights to weight the root mean square error index and the signal-to-noise ratio index to obtain the fusion index.

[0111] Step 3.7: When the weighted evaluation value is not within the preset range, reset the decomposition scale until the weighted evaluation value is within the preset range.

[0112] Step 4: Perform time domain conversion on the denoised vibration signal to obtain the frequency domain signal of the corresponding vibration signal;

[0113] In practical applications, the reflected wave is transmitted back to the concrete surface, and its signal is received by sensors pre-placed on the concrete surface. The reflected wave generates a new reflected wave on the concrete surface, which is then transmitted back to the interior of the concrete lining. This back-and-forth multiple reflections will make the waveform have obvious periodicity, which can be decomposed into multiple waves with different frequencies and peaks in the spectrum. Since the signal composed of multiple frequency components will interfere with or overlap with each other in the time domain, the waveform shape will be irregular or difficult to understand intuitively. Frequency domain analysis can decompose the signal into its frequency components, making the energy or amplitude of a specific frequency more clear and distinguishable, and providing important information about the signal characteristics. This paper uses fast Fourier transform to convert time domain data to obtain frequency domain information of the corresponding vibration signal.

[0114] Fourier transform is to decompose the signal into the linear superposition of the fundamental frequency and a series of harmonics, thereby converting the signal from the time domain to the frequency domain. There are two forms of Fourier transform: continuous Fourier transform (CFT) and discrete Fourier transform (DFT). Fast Fourier transform (FFT) is a commonly used signal transformation method and a method for quickly calculating DFT. Its basic idea is to decompose the calculation of DFT into multiple sub-problems, recursively solve these sub-problems, and finally merge the results. That is, by decomposing the DFT matrix into the product of sparse factors to quickly perform calculations, the computational complexity is reduced. The more sampling points are input, the more obvious the speed of FFT is. This divide-and-conquer idea greatly reduces the time complexity of the FFT algorithm. The original DFT time complexity is O(n 2 ), while FFT is only O(NlogN), so the calculation is done using the Fast Fourier Transform. The result of the Fast Fourier Transform is to transform the signal from the time domain to the frequency domain to highlight the characteristics of the disease. The calculation formula is as follows:

[0115]

[0116] Where c≥0, e v (c) = x(2c), o d (c) = x(2c+1), N is the total number of sampling points, k is the frequency index used to represent different frequency components, X k1 Represents the first N / 2 frequency components of the final Fourier transform result, Represents the last N / 2 frequency components of the Fourier transform result, k1 represents the frequency index, which determines the position of the frequency component in the frequency domain, e v (c) represents the even part, that is, the sampling points of all even index positions in signal c, x(2c) represents extracting the sample values ​​of the signal and rearranging them so that the indexes of these sample values ​​are even indexes, o d (c) represents the odd part, that is, the sampling points of all odd index positions in signal c, DFT represents the discrete Fourier transform function, represents the rotation factor.

[0117] Step 5: Complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

[0118] The amplitude of the periodic fluctuation phase in the time domain signal is extracted. When the amplitude is greater than the set threshold, a cavity appears at the corresponding point in the subway tunnel.

[0119] The peak value in the frequency domain signal is extracted. When the peak value exceeds the preset threshold, a hole appears at the corresponding point in the subway tunnel.

[0120] The vibration time domain data of each point measured in the test are as follows: Figure 4-7 As shown, the horizontal axis of the signal represents time, and the vertical axis represents acceleration value; Figure 4 、 Figure 6 They are the signals without holes in areas ② and ③, Figure 5 、 Figure 7 are the signals of the corresponding points where the holes are placed in regions ② and ③ respectively. From the local waveform, the amplitude in the periodic fluctuation stage is extracted and the Figure 4-7 The amplitude value (periodic vibration in the amplified area, the difference from the highest point to the lowest point in one cycle, the time of one cycle is 0.021 seconds). The results show that Figure 4-7 The amplitude values ​​are 0.116, 0.25, 0.052, and 0.58, respectively. Within a stable fluctuation period, the amplitude values ​​in areas containing voids are significantly greater than those in areas without voids. The value for the 1.5L void is similar to that for points without voids. The amplitude increases with increasing void volume, with the amplitude at points with continuous voids exceeding 10 times. The amplitude gradually decreases with increasing burial depth. In particular, at a burial depth of 30 cm, it is impossible to distinguish whether a void exists.

[0121] Fast Fourier transform is used to transform the time domain data to obtain the frequency domain information of the corresponding vibration signal. The results are as follows: Figure 8-11 shown. Figure 8 The circled part is the peak value of the entire signal. The maximum Y value here is the acceleration value, and the X-axis value corresponding to the maximum value is the frequency value.

[0122] The frequency peaks in the areas with voids are greater than those in the areas without voids, both occurring at 50.8 MHz. The amplitude increases with increasing simulated void size, most notably in the areas with pre-buried continuous voids, where the amplitude is an order of magnitude greater than that of the areas without voids. The amplitudes for simulated voids of the same size are relatively stable, ranging from approximately 0.002 for 5L voids, 0.005 for 12L voids, and 0.02 for continuous voids. The amplitudes for both groups without voids are approximately 0.001. Furthermore, the smaller value for the 1.5L void is not significantly different from the absence of a void, making it difficult to distinguish the presence of a 1.5L void. As the burial depth increases, the signal amplitude decreases, making it impossible to identify voids buried deeper than 16.5 cm. The amplitudes are relatively stable at the same burial depth, with an average value of 0.00186 for the 5L void at 0 cm.

[0123] The present invention can effectively improve the accuracy of cavity detection results and reduce the possibility of misjudgment by performing denoising processing and frequency domain analysis on the collected data.

[0124] The present invention also provides a system for detecting cavities behind subway shield tunnel walls, comprising:

[0125] Point marking module, used to mark the points to be inspected on the tube segments of subway tunnels with a marker pen;

[0126] A vibration signal acquisition module is used to strike a point and collect the vibration signal of the point using an acceleration sensor;

[0127] A denoising module is used to denoise the vibration signal of each point to obtain a denoised vibration signal;

[0128] A time domain conversion module is used to perform time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal;

[0129] The cavity detection module is used to complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

[0130] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for detecting voids behind the wall of a subway shield tunnel are implemented. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as the beneficial effects of the method for detecting voids behind the wall of a subway shield tunnel described in the above-mentioned technical solution, and will not be repeated here.

[0131] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technical solution that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for detecting cavities behind a subway shield tunnel wall, characterized in that: include: Step 1: Use a marker to mark the points to be inspected on the subway tunnel segments; Step 2: Hit the point with a rebound hammer and use an accelerometer to collect the vibration signal of the point; Step 3: De-noise the vibration signal at each point to obtain a denoised vibration signal; The step 3 comprises: Step 3.1: Set the decomposition scale and decompose the vibration signal of each point according to the set decomposition scale to obtain the wavelet coefficients; Step 3.2: Determine the denoising threshold based on the set decomposition scale; Step 3.3: When the wavelet coefficient is greater than or equal to the denoising threshold, the wavelet coefficient is processed using a processing function to generate processed wavelet coefficients; The step 3.3 includes: Using the formula: The wavelet coefficients are processed to generate the processed wavelet coefficients; where w′ i,k Represents the processed wavelet coefficients, sign represents the sign function, w i,k represents the kth original wavelet coefficient at the i-th decomposition scale, λ represents the denoising threshold, N represents the length of the vibration signal, and σ represents the noise standard deviation; Step 3.4: When the wavelet coefficient is less than the denoising threshold, the corresponding wavelet coefficient is changed to 0; Step 3.5: reconstructing the processed wavelet coefficients to form a denoised vibration signal; Step 3.6: Use fusion indicators to evaluate the denoised vibration signal; The step 3.6 includes: Step 3.6.1: Perform a preliminary evaluation of the denoised vibration signal to obtain the signal's root mean square error (RMS) and signal-to-noise ratio (SNR). In step 3.6.1, the calculation formulas for the root mean square error index and the signal-to-noise ratio index are: Where x(i) represents the original vibration signal, y(i) represents the signal-to-noise ratio, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the root mean square error; Step 3.6.2: Normalize the root mean square error index and the signal-to-noise ratio index to obtain the normalized root mean square error index and signal-to-noise ratio index; Step 3.6.3: Calculate the weights of the normalized root mean square error and signal-to-noise ratio indicators; The step 3.6.3 includes: Calculate the ratio of each indicator to all indicators and determine the weight based on the ratio; the weight calculation formula is: Among them, P ij It represents the ratio of the jth index when the decomposition scale is i, x ij It represents the value of the jth indicator when the decomposition layer is i after normalization, m represents the number of decomposition scales, and w j represents the weight of the j-th indicator; Step 3.6.4: Use the weights to weight the root mean square error index and the signal-to-noise ratio index to obtain the fusion index; Step 3.7: When the weighted evaluation value is not within the preset range, reset the decomposition scale until the weighted evaluation value is within the preset range; Step 4: Perform time domain conversion on the denoised vibration signal to obtain the frequency domain signal of the corresponding vibration signal; Step 5: Complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

2. The method for detecting cavities behind the wall of a subway shield tunnel according to claim 1, characterized in that: The step 4: performing time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal, including: Fast Fourier transform is used to transform the denoised vibration signal to obtain the frequency domain signal of the corresponding vibration signal; the conversion formula is: Where c≥0, e v (c) = x(2c), o d (c) = x(2c+1), N is the total number of sampling points, k is the frequency index used to represent different frequency components, X k1 Represents the first N / 2 frequency components of the final Fourier transform result, Represents the last N / 2 frequency components of the Fourier transform result, k1 represents the frequency index, which determines the position of the frequency component in the frequency domain, e v (c) represents the even part, that is, the sampling points of all even index positions in signal c, x(2c) represents extracting the sample values ​​of the signal and rearranging them so that the indexes of these sample values ​​are even indexes, o d (c) represents the odd part, that is, the sampling points of all odd index positions in signal c, DFT represents the discrete Fourier transform function, represents the rotation factor.

3. The method for detecting cavities behind the wall of a subway shield tunnel according to claim 2, characterized in that: The step 5: completing the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal, including: The amplitude of the periodic fluctuation phase in the time domain signal is extracted. When the amplitude is greater than the set threshold, a cavity appears at the corresponding point in the subway tunnel. The peak value in the frequency domain signal is extracted. When the peak value exceeds the preset threshold, a hole appears at the corresponding point in the subway tunnel.

4. A system for detecting cavities behind subway shield tunnel walls, characterized in that: include: Point marking module, used to mark the points to be inspected on the tube segments of subway tunnels with a marker pen; A vibration signal acquisition module is used to strike a point and collect the vibration signal of the point using an acceleration sensor; A denoising module is used to denoise the vibration signal of each point to obtain a denoised vibration signal; Among them, denoising processing includes: Step 3.1: Set the decomposition scale and decompose the vibration signal of each point according to the set decomposition scale to obtain the wavelet coefficients; Step 3.2: Determine the denoising threshold based on the set decomposition scale; Step 3.3: When the wavelet coefficient is greater than or equal to the denoising threshold, the wavelet coefficient is processed using a processing function to generate processed wavelet coefficients; The step 3.3 includes: Using the formula: The wavelet coefficients are processed to generate the processed wavelet coefficients; where w′ i,k Represents the processed wavelet coefficients, sign represents the sign function, w i,k represents the kth original wavelet coefficient at the i-th decomposition scale, λ represents the denoising threshold, N represents the length of the vibration signal, and σ represents the noise standard deviation; Step 3.4: When the wavelet coefficient is less than the denoising threshold, the corresponding wavelet coefficient is changed to 0; Step 3.5: reconstructing the processed wavelet coefficients to form a denoised vibration signal; Step 3.6: Use fusion indicators to evaluate the denoised vibration signal; The step 3.6 includes: Step 3.6.1: Perform a preliminary evaluation of the denoised vibration signal to obtain the signal's root mean square error (RMS) and signal-to-noise ratio (SNR). In step 3.6.1, the calculation formulas for the root mean square error index and the signal-to-noise ratio index are: Where x(i) represents the original vibration signal, y(i) represents the signal-to-noise ratio, n represents the sampling point number, N represents the sampling length, SNR represents the first evaluation value, and RMSE represents the root mean square error; Step 3.6.2: Normalize the root mean square error index and the signal-to-noise ratio index to obtain the normalized root mean square error index and signal-to-noise ratio index; Step 3.6.3: Calculate the weights of the normalized root mean square error and signal-to-noise ratio indicators; The step 3.6.3 includes: Calculate the ratio of each indicator to all indicators and determine the weight based on the ratio; the weight calculation formula is: Among them, P ij It represents the ratio of the jth index when the decomposition scale is i, x ij It represents the value of the jth indicator when the decomposition layer is i after normalization, m represents the number of decomposition scales, and w j represents the weight of the j-th indicator; Step 3.6.4: Use the weights to weight the root mean square error index and the signal-to-noise ratio index to obtain the fusion index; Step 3.7: When the weighted evaluation value is not within the preset range, reset the decomposition scale until the weighted evaluation value is within the preset range; A time domain conversion module is used to perform time domain conversion on the denoised vibration signal to obtain a frequency domain signal corresponding to the vibration signal; The cavity detection module is used to complete the detection of the cavity behind the subway shield tunnel wall based on the time domain signal of the denoised vibration signal and the frequency domain signal.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for detecting cavities behind a subway shield tunnel wall according to any one of claims 1 to 3 are implemented.

Images