Method and device for rapidly detecting shear wave velocity of high fill roadbed

Abstract

The invention relates to the field of wave velocity detection, and relates to a high fill roadbed shear wave velocity rapid detection method and device, the method comprises the following steps: obtaining vibration signals, the vibration signals comprising at least two modes of vibration signals; performing feature extraction on the vibration signal to obtain time-frequency feature map information; processing the time-frequency characteristic pattern information by using a wavelet transform algorithm to obtain separated signal information; calculating according to the time-frequency characteristic pattern information to obtain the first arrival time of the S wave; sending to an attention mechanism according to the separated signal information and the first arrival time of the S wave to obtain a feature vector after weighted fusion; the three-dimensional wave velocity field is inverted according to the feature vectors after weighted fusion, three-dimensional wave velocity field information is obtained, and the three-dimensional wave velocity field information is used for detecting the defects of the high-fill subgrade, the detection spatial resolution is improved to the millimeter level, and the requirement for high-precision evaluation of the layering compactness of the high-fill subgrade is met.

Description

Technical Field

[0001] This invention relates to the field of wave velocity detection, and more specifically, to a method and apparatus for rapid detection of shear wave velocity in high embankment subgrades. Background Technology

[0002] The evolution of shear wave velocity testing technology for high embankment subgrades is closely related to engineering needs. With the increasing demands for subgrade compaction and uniformity in large-scale projects such as highways and airports, traditional single-hole shear wave testing requires drilling to install sensors, causing localized damage to the subgrade, resulting in low construction efficiency and susceptibility to interference from groundwater levels and hard interlayers, making it difficult to capture compression waves in complex strata. Cross-hole shear wave testing requires multiple holes, which is costly and time-consuming, and drilling has a significant impact on the subgrade structure. Existing testing technologies, due to their destructive nature, low efficiency, and insufficient deep resolution, are gradually failing to meet the requirements for detecting the uniformity of subgrade compaction. Summary of the Invention

[0003] The purpose of this invention is to provide a method and apparatus for rapid detection of shear wave velocity in high embankment roadbeds, so as to improve the above-mentioned problems.

[0004] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0005] On the one hand, embodiments of this application provide a method for rapid detection of shear wave velocity in high embankment subgrades, the method comprising:

[0006] Acquire vibration signals, the vibration signals including vibration signals of at least two modes;

[0007] Feature extraction is performed on the vibration signal to obtain time-frequency feature map information;

[0008] The time-frequency feature map information is processed using a wavelet transform algorithm to obtain the separated signal information;

[0009] The first arrival time of the S-wave is calculated based on the time-frequency characteristic map information.

[0010] Based on the separated signal information and the first arrival time of the S-wave, the signal is sent to the attention mechanism to obtain a weighted fused feature vector;

[0011] The three-dimensional wave velocity field is inverted based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

[0012] Secondly, embodiments of this application provide a rapid detection device for shear wave velocity in high embankment subgrades, the device comprising:

[0013] An acquisition module is used to acquire vibration signals, the vibration signals including vibration signals of at least two modes;

[0014] The first processing module is used to extract features from the vibration signal to obtain time-frequency feature map information;

[0015] The second processing module is used to process the time-frequency feature map information using a wavelet transform algorithm to obtain the separated signal information;

[0016] The third processing module is used to calculate the first arrival time of the S-wave based on the time-frequency feature map information.

[0017] The fourth processing module is used to send the separated signal information and the first arrival time of the S wave to the attention mechanism to obtain the weighted fused feature vector;

[0018] The fifth processing module is used to invert the three-dimensional wave velocity field based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

[0019] Thirdly, embodiments of this application provide an apparatus, which includes a memory and a processor. The memory stores a computer program; the processor executes the computer program to implement the steps of the above-described rapid detection method for shear wave velocity in high embankment subgrades.

[0020] Fourthly, embodiments of this application provide a readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for rapid detection of shear wave velocity in high embankment subgrades.

[0021] The beneficial effects of this invention are as follows:

[0022] This invention uses a multi-source excitation device to replace the traditional drilling triggering method, fundamentally avoiding the problem of "hole wall soil disturbance" caused by drilling operations. This effectively reduces the risk of roadbed settlement induced by seepage channels formed by drilling. In the signal acquisition and processing stage, a multi-mode excitation source synchronously excites shear waves and surface waves, and a high-density sensor array is used to collect vibration signals. Time-frequency feature map information is generated by feature extraction algorithm to accurately capture the first arrival time of S-wave. Subsequently, wavelet transform algorithm is used to process the time-frequency feature map to separate the signal. Finally, an attention mechanism is introduced to realize the rapid detection of shear wave velocity in high-fill roadbeds, improving the detection spatial resolution to the millimeter level and meeting the high-precision assessment requirements of layered compaction of high-fill roadbeds.

[0023] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the rapid detection method for shear wave velocity in high embankment subgrades as described in an embodiment of the present invention.

[0026] Figure 2 This is a schematic diagram of the rapid detection device for shear wave velocity in high embankment subgrade as described in an embodiment of the present invention.

[0027] Figure 3 This is a schematic diagram of the rapid detection device for shear wave velocity in high embankment subgrade as described in an embodiment of the present invention.

[0028] The diagram is labeled as follows: 800, Rapid detection device for shear wave velocity of high embankment subgrade; 801, Processor; 802, Memory; 803, Multimedia component; 804, I / O interface; 805, Communication component; 901, Acquisition module; 902, First processing module; 903, Second processing module; 904, Third processing module; 905, Fourth processing module; 906, Fifth processing module. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0030] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0031] Example 1:

[0032] This embodiment provides a rapid detection method for shear wave velocity in high embankment subgrades. It can be understood that this embodiment can be used to construct a scenario, such as a highway construction project in a plateau mountainous area where the compaction degree of a section of high embankment subgrade needs to be tested. The geological conditions in this area are complex, with loose gravel and soil interlayers and seasonal groundwater level fluctuations. Traditional borehole testing methods are prone to causing slope instability risks, and the quality of deep compaction is difficult to accurately assess.

[0033] See Figure 1 The figure shows that the method includes steps S1, S2, S3, S4, S5 and S6.

[0034] Step S1: Acquire vibration signals, wherein the vibration signals include vibration signals of at least two modes;

[0035] Step S1 further includes steps S11, S12, S13, and S14, which specifically include:

[0036] Step S11: Obtain the detection depth threshold;

[0037] Step S12: Determine whether the detection depth is greater than the detection depth threshold, and obtain the determination result;

[0038] Step S13: Determine the type of excitation source based on the judgment result;

[0039] Step S14: Obtain vibration signals of different modes according to the type of excitation source.

[0040] In this embodiment, the excitation source types include an electromagnetic vibrator and a pneumatic impact device. The electromagnetic vibrator uses the principle of electromagnetic induction to generate a continuous sinusoidal vibration signal by controlling the current frequency (0.1-100Hz). It is mainly suitable for shallow high-resolution detection and covers the penetration requirements of different filler particle sizes (5-80mm) through frequency conversion control. The pneumatic impact device uses compressed air (pressure range 0.5-1MPa) to drive the impact hammer and generate high-frequency pulse vibration (100-500Hz). Its penetration depth can reach 30m and is used for deep shear wave excitation. In this embodiment, the excitation source is automatically switched based on the preset detection depth. When the detection depth is less than 5m, the electromagnetic vibrator is used; when the detection depth is greater than 5m, the pneumatic impact device is used.

[0041] Step S2: Extract features from the vibration signal to obtain time-frequency feature map information;

[0042] Step S2 further includes steps S21, S22, and S23, which specifically include:

[0043] Step S21: Preprocess the vibration signal to obtain a preprocessed vibration signal;

[0044] In this step, preprocessing includes high-pass filtering and normalization.

[0045] Step S22: Perform a short-time Fourier transform on the preprocessed vibration signal to obtain the time-frequency energy spectrum;

[0046] In this step, the time-frequency energy spectrum of the signal is calculated. The window length is adaptively adjusted according to the signal frequency. For example, a longer window is used for low-frequency signals, and a shorter window is used for high-frequency signals. The specific calculation process for the adaptive window length is as follows:

[0047] ;

[0048] In the above formula, This indicates the window length, which controls the time resolution (number of sample points) of time-frequency analysis. The larger the window, the higher the frequency resolution but the lower the time resolution. Represents an empirical constant; This represents the local dominant frequency of the signal, obtained through short-time spectrum analysis.

[0049] It should be noted that in high embankment subgrades, different excitation sources (such as electromagnetic vibrators and pneumatic impact devices) will generate broadband signals (0.1-500Hz). The adaptive window can dynamically adjust the analysis scale for different frequency bands of signals such as shear waves (10-50Hz) and surface waves (5-30Hz), which helps to automatically pick up the first arrival time of S-waves through subsequent wavelet transform noise reduction and CNN feature recognition. At the same time, using a shorter window length for high-frequency signals can effectively focus on the abrupt changes of the signal in a short time, improve the time resolution, and accurately locate the first arrival time of high-frequency components such as shear waves (S-waves) and surface waves (R-waves). Using a longer window length for low-frequency signals can capture the periodic changes of the signal in the long time domain, improve the frequency resolution, and avoid missing low-frequency components due to excessively short windows.

[0050] Step S23: Send the time-frequency energy spectrum to a lightweight convolutional neural network to obtain time-frequency feature map information.

[0051] In this step, the lightweight convolutional neural network structure consists of 3 convolutional layers + pooling layers, filter size [32, 64, 128], kernel size 3×3, stride 1, and is used to extract time-frequency feature maps of S-waves, R-waves and other waveforms in the signal.

[0052] Step S3: Process the time-frequency feature map information using the wavelet transform algorithm to obtain the separated signal information;

[0053] Step S3 further includes steps S31, S32, S33, and S34, which specifically include:

[0054] Step S31: Decompose the time-frequency feature map information based on the preset wavelet basis function to obtain the frequency band signals corresponding to different frequency bands;

[0055] In this step, Daubechies wavelet is selected as the basis function, and the number of decomposition layers is N. The signal is decomposed into different frequency bands, specifically divided into the 4th frequency band: 10-20Hz (main S-wave frequency band); the 5th frequency band: 5-10Hz (low R-wave frequency band); and the 3rd frequency band: 20-40 Hz (high R-wave frequency band).

[0056] Step S32: Calculate the standard deviation of the signal in each frequency band;

[0057] In this step, the specific calculation process for the standard deviation is as follows:

[0058] ;

[0059] In the above formula, The noise standard deviation represents the detail coefficients of the j-th layer; This represents the detail coefficients of the j-th layer; the coefficient 0.6745 is an empirical value (based on Gaussian distribution characteristics). This represents the median function.

[0060] Step S33: Calculate soft threshold information;

[0061] In this step, the specific calculation process for the soft threshold information is as follows:

[0062] ;

[0063] In the above formula, This represents the threshold value for the j-th layer (units consistent with the signal, such as m / s²). Indicates the signal length (number of sample points). This represents the noise standard deviation of the detail coefficients at the j-th layer.

[0064] Step S34: Based on the soft threshold information and the standard deviation of each frequency band signal, reconstruct the frequency band signals corresponding to different frequency bands to obtain the separated signal information.

[0065] In this step, the processing using soft threshold information specifically includes:

[0066] ;

[0067] In the above formula, Represents the detail coefficients after filtering; It is a symbolic function; It represents soft threshold information, "shrinking" coefficients above the threshold instead of setting them directly to zero, preserving the phase and amplitude trends of the signal, avoiding abrupt changes or loss of features at the waveform edges caused by hard thresholding, preserving the steep leading edge of the shear wave (S-wave) and the low-frequency attenuation characteristics of the surface wave (R-wave), and ensuring the integrity of the waveform polarity and energy distribution in the time-frequency feature diagram.

[0068] The signal reconstruction process is as follows:

[0069] ;

[0070] In the above formula, The signal is separated; j represents the number of decomposition layers; k represents the time shift parameter. Indicates the coefficient length of each layer; Represents the detail coefficients after filtering; Represents the approximation coefficients of the Nth layer; Describe the wavelet basis functions; The scaling function is used to indicate that the separated S-wave signal is reconstructed from layers 4-5, and the separated R-wave signal is reconstructed from layers 3-4.

[0071] Step S4: Calculate the first arrival time of the S-wave based on the time-frequency characteristic map information;

[0072] In this step, the energy spectrum of the time-frequency feature map is calculated; then, the gradient of the energy spectrum is calculated along the time axis to obtain the energy gradient; based on the energy gradient, PTO (first arrival time) detection is performed, specifically: when The time at which the first arrival is determined is then considered the arrival time. It should be noted that... Represents the energy gradient; This represents the normalization threshold coefficient, which is typically set to 80%. This represents the global maximum value of the energy gradient. The output of this step includes the PTO timestamps for S-wave and R-wave: the time point at which the energy gradient was detected to exceed the threshold and the frequency band characteristics near the corresponding PTO time.

[0073] Step S5: Based on the separated signal information and the first arrival time of the S-wave, the signal is sent to the attention mechanism to obtain the weighted fused feature vector;

[0074] Step S5 further includes steps S51, S52, S53, S54, and S55, which specifically include:

[0075] Step S51: Obtain the strain field feature vector;

[0076] Step S52: Map the separated signal information and the first arrival time of the S wave to a unified embedding space to obtain the first embedding vector and the second embedding vector;

[0077] Step S53: Concatenate the strain field feature vector, the first embedding vector, and the second embedding vector to obtain the concatenated vector;

[0078] In this step, the concatenated vector is represented as: ,in, Represents the second embedding vector. Represents the first embedding vector. Let D represent the strain field feature vector, where D represents the dimension and N represents the length of the input feature sequence.

[0079] Step S54: Calculate the self-attention weight matrix based on the concatenated vector;

[0080] In this step, calculating the self-attention weight matrix is ​​a technique well-known to those skilled in the art, and therefore will not be described in detail here.

[0081] Step S55: Calculate the weighted fused feature vector based on the self-attention weight matrix.

[0082] In this step, the output of the self-attention weight matrix is ​​segmented according to the modal dimension, and the normalized weights are calculated. The specific process is as follows:

[0083] ;

[0084] in: The weights of mode i are represented by ( ); This means calculating the average by row. Matrix compression The vector is then divided into three scalar weights according to modality, thus outputting a weighted fused feature vector.

[0085] Step S6: Invert the three-dimensional wave velocity field based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

[0086] Step S6 further includes steps S61, S62, S63, and S64, which specifically include:

[0087] Step S61: Discretize the monitoring area into a material point grid to obtain a discretized three-dimensional material point grid model;

[0088] In this step, a specific implementation method is to discretize the monitoring area into a 20×20×6 material point grid (0.5m resolution) and initialize the wave velocity field to 1500m / s.

[0089] Step S62: Simulate the theoretical first arrival times of S-waves and R-waves based on the wave equation in a discretized three-dimensional material point mesh model;

[0090] In this step, the theoretical first arrival times of the S-wave and R-wave are calculated using forward modeling based on the wave equation. This forward modeling based on the wave equation includes:

[0091] ① Elastic wave equation (three-dimensional):

[0092] ;

[0093] In the above formula, Indicates the density of the medium (unit: kg / m³). Displacement vector (unit: m); Represents the stress tensor (unit: Pa); This indicates the external force source (unit: N / m³).

[0094] ② Discretized form (explicit time integral):

[0095] ;

[0096] In the above formula, Represents the displacement vector at time n; Represented as a time step, It is expressed as a time step (unit: s) and must satisfy stability conditions (such as the Courant-Friedrichs-Lewy condition).

[0097] ③ Stress calculation (based on displacement gradient):

[0098] ;

[0099] In the above formula, , Let λ represent the Lamé parameter, where λ is related to volumetric deformation and reflects the medium's ability to resist volumetric compression; μ is the shear modulus and reflects the medium's ability to resist shape deformation. Represents the components of the strain tensor; Indicates volumetric strain; This represents the Kronecker symbol, which is 1 when i=j and 0 otherwise.

[0100] ④ Boundary condition handling: Use absorbing boundary conditions (such as a perfectly matched layer PML) to reduce reflection interference.

[0101] It should be noted that, through explicit time integration and Lamé parameter modeling, the wave velocity differences in different subgrade materials (such as soil-rock mixtures and compacted clay) are accurately simulated, supporting fine inversion of the three-dimensional wave velocity field within a depth of 20m (resolution 0.5m×0.5m×1m). The absorbing boundary condition (PML) effectively suppresses reflection noise caused by artificially truncated boundaries, ensuring that the energy attenuation characteristics of the simulated wave field are consistent with the actual subgrade medium. Simultaneously, a three-dimensional wave velocity model within a depth of 20m is established through wave equation constraints. This effectively solves the problem of deep, unevenly compacted areas caused by "disordered filling sequence".

[0102] Step S63: Construct an error function based on the theoretical first arrival times of S-waves and R-waves and the weighted fused eigenvectors;

[0103] In this step, the error function constructed is specifically as follows:

[0104] ;

[0105] In the above formula, , and Modal weights dynamically assigned to the attention mechanism; and These represent the theoretical first arrival time of the S-wave and the first arrival time of the S-wave as identified by CNN, respectively. and These represent the theoretical first arrival time of the R-wave and the first arrival time of the R-wave as identified by the CNN, respectively. This represents the simulated value and the observed value of the strain rate.

[0106] Step S64: Iteratively update the wave velocity field using the L-BFGS optimization algorithm, minimize the error function, and obtain the three-dimensional wave velocity field information.

[0107] In this step, the iterative process is as follows:

[0108] Step a. Initialization: Set the initial wave velocity field parameter vector Initialize the approximate Hessian matrix (Identity matrix, to avoid Hessian singularity); Set the iteration counter k=0, and the maximum number of iterations K. max Convergence threshold Parameters such as these.

[0109] Step b. Calculate the objective function and gradient: based on the current wave velocity field. Through forward modeling calculations, the theoretical time... Calculate the error function value Calculate the gradient .

[0110] Step c. Constructing the search direction: Using the two-loop recursive algorithm of L-BFGS, the Hessian inverse matrix is ​​approximately calculated. The final search direction is:

[0111] Step d. Determine the step size using line search: Determine the optimal step size using a line search algorithm. This allows the objective function to decrease sufficiently along the search direction.

[0112] Step e. Parameter update: ;

[0113] Step f. Convergence check: If or ,or If the iteration terminates, the optimal value is output. Otherwise, proceed to the next iteration: This continues until the convergence principle is met.

[0114] Understandable, Indicates the norm threshold; This represents the norm of the gradient of the objective function. The iteration stops when any of the three convergence conditions are met, and the inverted three-dimensional wave velocity field distribution is output. When the measured wave velocity value is less than the critical wave velocity value, the wave velocity anomaly region is identified as a potential defect region and marked as a dark region.

[0115] It should be noted that gradient information is used to construct the search direction during the iteration process, and the optimal step size is determined by line search. It usually converges in 10-20 iterations (traditional gradient descent requires hundreds of iterations), which is suitable for rapid detection scenarios of high embankment subgrades (such as dynamic monitoring during construction).

[0116] This invention addresses the core pain points of traditional high-fill roadbed detection—namely, inaccuracy, slowness, and instability—by combining multimodal wavefield fusion data complementarity, an adaptive attention mechanism, and the L-BFGS optimization algorithm.

[0117] Example 2:

[0118] like Figure 2 As shown, this embodiment provides a rapid detection device for shear wave velocity in high embankment subgrades. The device includes an acquisition module 901, a first processing module 902, a second processing module 903, a third processing module 904, a fourth processing module 905, and a fifth processing module 906, specifically comprising:

[0119] Acquisition module 901 is used to acquire vibration signals, the vibration signals including vibration signals of at least two modes;

[0120] The first processing module 902 is used to extract features from the vibration signal to obtain time-frequency feature map information;

[0121] The second processing module 903 is used to process the time-frequency feature map information using a wavelet transform algorithm to obtain the separated signal information;

[0122] The third processing module 904 is used to calculate the first arrival time of the S-wave based on the time-frequency feature map information.

[0123] The fourth processing module 905 is used to send the separated signal information and the first arrival time of the S wave to the attention mechanism to obtain the weighted fused feature vector;

[0124] The fifth processing module 906 is used to invert the three-dimensional wave velocity field based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

[0125] In one specific embodiment of this disclosure, the acquisition module further includes a first acquisition unit, a judgment unit, a first processing unit, and a second processing unit, specifically including:

[0126] The first acquisition unit is used to acquire the detection depth threshold;

[0127] The judgment unit is used to determine whether the detection depth is greater than the detection depth threshold and to obtain a judgment result.

[0128] The first processing unit is used to determine the type of excitation source based on the judgment result;

[0129] The second processing unit is used to acquire vibration signals of different modes according to the type of excitation source.

[0130] In one specific embodiment of this disclosure, the first processing module further includes a third processing unit, a fourth processing unit, and a fifth processing unit, specifically including:

[0131] The third processing unit is used to preprocess the vibration signal to obtain the preprocessed vibration signal;

[0132] The fourth processing unit is used to perform a short-time Fourier transform on the preprocessed vibration signal to obtain the time-frequency energy spectrum.

[0133] The fifth processing unit is used to send the time-frequency energy spectrum to a lightweight convolutional neural network to obtain time-frequency feature map information.

[0134] In one specific embodiment of this disclosure, the second processing module further includes a sixth processing unit, a first computing unit, a second computing unit, and a seventh processing unit, specifically including:

[0135] The sixth processing unit is used to decompose the time-frequency feature map information based on the preset wavelet basis function to obtain the frequency band signals corresponding to different frequency bands;

[0136] The first calculation unit is used to calculate the standard deviation of the signal for each frequency band.

[0137] The second calculation unit is used to calculate soft threshold information;

[0138] The seventh processing unit is used to reconstruct the frequency band signals corresponding to different frequency bands based on the soft threshold information and the standard deviation of each frequency band signal to obtain the separated signal information.

[0139] In one specific embodiment of this disclosure, the fourth processing module further includes a second acquisition unit, an eighth processing unit, a ninth processing unit, a tenth processing unit, and an eleventh processing unit, specifically including:

[0140] The second acquisition unit is used to acquire the strain field feature vector;

[0141] The eighth processing unit is used to map the separated signal information and the first arrival time of the S-wave to a unified embedding space to obtain a first embedding vector and a second embedding vector.

[0142] The ninth processing unit is used to concatenate the strain field feature vector, the first embedding vector, and the second embedding vector to obtain the concatenated vector;

[0143] The tenth processing unit is used to calculate the self-attention weight matrix based on the concatenated vector;

[0144] The eleventh processing unit is used to calculate the weighted fused feature vector based on the self-attention weight matrix.

[0145] In one specific embodiment of this disclosure, the fifth processing module further includes a twelfth processing unit, a thirteenth processing unit, a fourteenth processing unit, and a fifteenth processing unit, specifically comprising:

[0146] The twelfth processing unit is used to discretize the monitoring area into a material point grid to obtain a discretized three-dimensional material point grid model.

[0147] The thirteenth processing unit is used to simulate the theoretical first arrival times of S-waves and R-waves based on the wave equation in a discretized three-dimensional material point grid model.

[0148] The fourteenth processing unit is used to construct an error function based on the theoretical first arrival times of S-waves and R-waves and the weighted fused eigenvectors.

[0149] The fifteenth processing unit is used to iteratively update the wave velocity field using the L-BFGS optimization algorithm, minimize the error function, and obtain the three-dimensional wave velocity field information.

[0150] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0151] Example 3:

[0152] Corresponding to the above method embodiments, this embodiment also provides a rapid detection device for shear wave velocity of high embankment subgrade. The rapid detection device for shear wave velocity of high embankment subgrade described below and the rapid detection method for shear wave velocity of high embankment subgrade described above can be referred to in correspondence with each other.

[0153] Figure 3 This is a block diagram illustrating a rapid detection device 800 for shear wave velocity in high embankment subgrades, according to an exemplary embodiment. Figure 3 As shown, the rapid detection device 800 for shear wave velocity of high embankment subgrade may include: a processor 801 and a memory 802. The rapid detection device 800 for shear wave velocity of high embankment subgrade may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0154] The processor 801 controls the overall operation of the high embankment subgrade shear wave velocity rapid detection device 800 to complete all or part of the steps in the aforementioned high embankment subgrade shear wave velocity rapid detection method. The memory 802 stores various types of data to support the operation of the high embankment subgrade shear wave velocity rapid detection device 800. This data may include, for example, instructions for any application or method operating on the high embankment subgrade shear wave velocity rapid detection device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the high embankment subgrade shear wave velocity rapid detection device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, and an NFC module.

[0155] In an exemplary embodiment, the high embankment subgrade shear wave velocity rapid detection device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described high embankment subgrade shear wave velocity rapid detection method.

[0156] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the above-described rapid detection method for shear wave velocity of high embankment subgrades. For example, the computer-readable storage medium may be the memory 802 including the program instructions, which may be executed by the processor 801 of the rapid detection device 800 for shear wave velocity of high embankment subgrades to complete the above-described rapid detection method for shear wave velocity of high embankment subgrades.

[0157] Example 4:

[0158] Corresponding to the above method embodiments, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the rapid detection method for shear wave velocity of high embankment subgrade described above.

[0159] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the rapid detection method for shear wave velocity in high embankment subgrades described in the above method embodiments.

[0160] Specifically, the readable storage medium can be a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other readable storage medium capable of storing program code.

[0161] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0162] 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 variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A rapid detection method for shear wave velocity in high embankment subgrades, characterized in that, include: Acquire vibration signals, the vibration signals including vibration signals of at least two modes; Feature extraction is performed on the vibration signal to obtain time-frequency feature map information; The time-frequency feature map information is processed using a wavelet transform algorithm to obtain the separated signal information; The first arrival time of the S-wave is calculated based on the time-frequency characteristic map information. Based on the separated signal information and the first arrival time of the S-wave, the signal is sent to the attention mechanism to obtain a weighted fused feature vector; The three-dimensional wave velocity field is inverted based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

2. The method for rapid detection of shear wave velocity in high embankment subgrades according to claim 1, characterized in that, Acquiring vibration signals, including: Obtain the detection depth threshold; Determine whether the detection depth is greater than the detection depth threshold to obtain the determination result; The type of excitation source is determined based on the judgment result; Vibration signals of different modes are obtained according to the type of excitation source.

3. The rapid detection method for shear wave velocity in high embankment subgrades according to claim 1, characterized in that, Feature extraction is performed on the vibration signal to obtain time-frequency feature map information, including: The vibration signal is preprocessed to obtain a preprocessed vibration signal; The preprocessed vibration signal is subjected to a short-time Fourier transform to obtain the time-frequency energy spectrum; The time-frequency energy spectrum is sent to a lightweight convolutional neural network to obtain time-frequency feature map information.

4. The rapid detection method for shear wave velocity in high embankment subgrades according to claim 1, characterized in that, The time-frequency feature map information is processed using a wavelet transform algorithm, including: The time-frequency feature map information is decomposed based on the preset wavelet basis function to obtain the frequency band signals corresponding to different frequency bands; Calculate the standard deviation of the signal for each frequency band; Calculate soft threshold information; Based on the soft threshold information and the standard deviation of each frequency band signal, the frequency band signals corresponding to different frequency bands are reconstructed to obtain the separated signal information.

5. The rapid detection method for shear wave velocity in high embankment subgrades according to claim 1, characterized in that, Based on the separated signal information and the first arrival time of the S-wave, the signal is sent to the attention mechanism, including: Obtain the strain field feature vector; The separated signal information and the first arrival time of the S-wave are mapped to a unified embedding space to obtain a first embedding vector and a second embedding vector. The strain field feature vector, the first embedding vector, and the second embedding vector are concatenated to obtain the concatenated vector. Calculate the self-attention weight matrix based on the concatenated vector; The weighted fused feature vector is calculated based on the self-attention weight matrix.

6. A rapid detection device for shear wave velocity in high embankment roadbeds, characterized in that, include: An acquisition module is used to acquire vibration signals, the vibration signals including vibration signals of at least two modes; The first processing module is used to extract features from the vibration signal to obtain time-frequency feature map information; The second processing module is used to process the time-frequency feature map information using a wavelet transform algorithm to obtain the separated signal information; The third processing module is used to calculate the first arrival time of the S-wave based on the time-frequency feature map information. The fourth processing module is used to send the separated signal information and the first arrival time of the S wave to the attention mechanism to obtain the weighted fused feature vector; The fifth processing module is used to invert the three-dimensional wave velocity field based on the weighted and fused feature vector to obtain three-dimensional wave velocity field information, which is used to detect defects in high embankment subgrade.

7. The rapid detection device for shear wave velocity of high embankment subgrade according to claim 6, characterized in that, The acquisition module includes: The first acquisition unit is used to acquire the detection depth threshold; The judgment unit is used to determine whether the detection depth is greater than the detection depth threshold and to obtain a judgment result. The first processing unit is used to determine the type of excitation source based on the judgment result; The second processing unit is used to acquire vibration signals of different modes according to the type of excitation source.

8. The rapid detection device for shear wave velocity of high embankment subgrade according to claim 6, characterized in that, The first processing module includes: The third processing unit is used to preprocess the vibration signal to obtain the preprocessed vibration signal; The fourth processing unit is used to perform a short-time Fourier transform on the preprocessed vibration signal to obtain the time-frequency energy spectrum. The fifth processing unit is used to send the time-frequency energy spectrum to a lightweight convolutional neural network to obtain time-frequency feature map information.

9. The rapid detection device for shear wave velocity of high embankment subgrade according to claim 6, characterized in that, The second processing module includes: The sixth processing unit is used to decompose the time-frequency feature map information based on the preset wavelet basis function to obtain the frequency band signals corresponding to different frequency bands; The first calculation unit is used to calculate the standard deviation of the signal for each frequency band. The second calculation unit is used to calculate soft threshold information; The seventh processing unit is used to reconstruct the frequency band signals corresponding to different frequency bands based on the soft threshold information and the standard deviation of each frequency band signal to obtain the separated signal information.

10. The rapid detection device for shear wave velocity of high embankment subgrade according to claim 6, characterized in that, The fourth processing module includes: The second acquisition unit is used to acquire the strain field feature vector; The eighth processing unit is used to map the separated signal information and the first arrival time of the S-wave to a unified embedding space to obtain a first embedding vector and a second embedding vector. The ninth processing unit is used to concatenate the strain field feature vector, the first embedding vector, and the second embedding vector to obtain the concatenated vector; The tenth processing unit is used to calculate the self-attention weight matrix based on the concatenated vector; The eleventh processing unit is used to calculate the weighted fused feature vector based on the self-attention weight matrix.

Images