Tunnel structure detection method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a tunnel structure detection method and device, computer equipment, a readable storage medium and a program product. The method comprises the following steps: determining a key abnormal section in a tunnel structure; based on a preset strain cloud picture generation model and the strain monitoring data corresponding to the key abnormal section, determining a first strain cloud picture of the key abnormal section in an abnormal state; comparing the first strain cloud picture with a second strain cloud picture in a normal state to obtain a strain cloud picture evolution process and a deformation degree corresponding to the key abnormal section; and determining a recovery strategy based on the evolution process of the strain cloud picture and the deformation degree. By adopting the method, the recognition accuracy of the deformed section caused by abnormal disturbance can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of operation monitoring and detection of heavy-haul railway tunnels, and particularly to a tunnel structure detection method, device, computer device, readable storage medium, and program product. Background Art

[0002] During the service period of heavy-haul railways, due to the large structural deformation caused by trains with heavy loads passing through tunnels, once there is an abnormal disturbance source in the outside world, it is very likely to cause irreversible damage to railway tunnels and trains. In related technologies, Brillouin Optical Time Domain Analysis (BOTDA) distributed optical fibers are usually used to monitor the strain of tunnels, which can only extract the structural strain of tunnel sections, and the recognition accuracy of the deformation sections caused by abnormal disturbances is relatively low. Summary of the Invention

[0003] Based on this, it is necessary to provide a tunnel structure detection method, device, computer device, readable storage medium, and program product that can improve the recognition accuracy of deformation sections caused by abnormal disturbances for the above technical problems.

[0004] In a first aspect, the present application provides a tunnel structure detection method, including:

[0005] Determine the key abnormal sections in the tunnel structure; and based on a preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal sections, determine the first strain cloud map of the key abnormal sections in an abnormal state;

[0006] Compare the first strain cloud map with the second strain cloud map in a normal state to obtain the strain cloud map evolution process and deformation degree corresponding to the key abnormal sections; determine a recovery strategy based on the strain cloud map evolution process and the deformation degree.

[0007] In one of the embodiments, the method further includes:

[0008] Obtain the tunnel structure and determine the abnormal vibration signal characteristics of the tunnel structure;

[0009] Input the abnormal vibration signal characteristics into a preset vibration event classification model to determine the type of vibration event corresponding to the abnormal vibration signal, and based on the type of vibration event, determine the risk level of the tunnel structure;

[0010] When the risk level meets a preset abnormal condition, execute the step of determining the key abnormal sections in the tunnel structure.

[0011] In one embodiment, the obtaining of the tunnel structure and determining the characteristics of the abnormal vibration signal of the structure includes:

[0012] Obtain the vibration signal of the tunnel structure; perform noise reduction processing on the vibration signal based on a preset spectral subtraction method to obtain the vibration signal after noise reduction processing; determine an endpoint detection threshold based on the vibration signal after noise reduction processing;

[0013] Perform abnormal screening processing on the vibration signal after noise reduction processing based on the endpoint detection threshold to obtain an abnormal vibration signal;

[0014] Extract features from the abnormal vibration signal based on the local characteristic scale decomposition method combined with multi-scale permutation entropy to obtain the characteristics of the abnormal vibration signal.

[0015] In one embodiment, the determining of the first strain cloud map of the key abnormal section in the abnormal state based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section includes:

[0016] Determine the strain data vector corresponding to the key abnormal section based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section;

[0017] Restore the strain data vector to the three-dimensional space where the strain monitoring data is located to obtain a target strain data vector, and perform surface fitting processing on the target strain data vector based on a preset surface fitting algorithm to obtain a non-uniform surface;

[0018] Perform reconstruction processing on the non-uniform surface based on the surface network segmentation energy algorithm to obtain the first strain cloud map of the key abnormal section in the abnormal state.

[0019] In one embodiment, the performing of the reconstruction processing on the non-uniform surface based on the surface network segmentation energy algorithm to obtain the first strain cloud map of the key abnormal section in the abnormal state includes:

[0020] Construct an original surface based on the target strain data vector, perform surface construction processing on the non-uniform surface based on local bicubic surface interpolation to obtain a plurality of bicubic Bezier surface patches, and splice the plurality of bicubic Bezier surface patches to obtain an auxiliary surface;

[0021] Perform rough fitting on the auxiliary surface based on a preset fairing threshold to obtain a fairing region and an unfairing region;

[0022] Construct grid lines in the unfairing region, remove the bicubic Bezier surface patches corresponding to the unfairing region, and perform local fairing processing on the grid lines corresponding to the unfairing region to obtain target grid lines;

[0023] Adjust the target grid line based on the normal distance between the auxiliary surface and the original surface to obtain an adjusted target grid line;

[0024] Determine the main curve family for the grid lines in the first target direction of the adjusted target grid line, and determine the guiding curve family for the grid lines in the second target direction of the adjusted target grid line; optimize the auxiliary surface based on the main curve family and the guiding curve family to obtain a target auxiliary surface, where the boundary of the target auxiliary surface conforms to the first-order continuity condition with the boundary of the original surface;

[0025] Fit the target strain data vector and the target auxiliary surface to obtain a first strain nephogram.

[0026] In one embodiment, the method further includes:

[0027] Determine a sample section of the tunnel structure, and obtain initial sample strain nephogram data and initial sample strain monitoring data in a normal state;

[0028] Normalize the initial sample strain nephogram data corresponding to the sample section to obtain sample strain nephogram data; standardize the initial sample strain monitoring data in the sample section to obtain target sample strain monitoring data;

[0029] Perform coordinate transformation on the sample strain nephogram data based on a preset transformation algorithm to obtain target sample strain nephogram data;

[0030] Input the target sample strain monitoring data into an initial strain nephogram generation model to obtain predicted strain nephogram data, and correct the initial strain nephogram generation model based on the predicted strain nephogram data and the target sample strain nephogram data to obtain a preset strain nephogram generation model.

[0031] In a second aspect, the present application further provides a tunnel structure detection device, including:

[0032] A determination module, configured to determine key abnormal sections in the tunnel structure; and determine a first strain nephogram of the key abnormal sections in an abnormal state based on a preset strain nephogram generation model and the strain monitoring data corresponding to the key abnormal sections;

[0033] A detection module, configured to compare the first strain nephogram with a second strain nephogram in a normal state to obtain the strain nephogram evolution process and deformation degree corresponding to the key abnormal sections; determine a recovery strategy based on the strain nephogram evolution process and the deformation degree.

[0034] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0035] Determine a key abnormal section in the tunnel structure; and based on a preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine a first strain cloud map of the key abnormal section in an abnormal state;

[0036] Compare the first strain cloud map with a second strain cloud map in a normal state to obtain the evolution process and deformation degree of the strain cloud map corresponding to the key abnormal section; determine a recovery strategy based on the evolution process of the strain cloud map and the deformation degree.

[0037] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0038] Determine a key abnormal section in the tunnel structure; and based on a preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine a first strain cloud map of the key abnormal section in an abnormal state;

[0039] Compare the first strain cloud map with a second strain cloud map in a normal state to obtain the evolution process and deformation degree of the strain cloud map corresponding to the key abnormal section; determine a recovery strategy based on the evolution process of the strain cloud map and the deformation degree.

[0040] In a fifth aspect, the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0041] Determine a key abnormal section in the tunnel structure; and based on a preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine a first strain cloud map of the key abnormal section in an abnormal state;

[0042] Compare the first strain cloud map with a second strain cloud map in a normal state to obtain the evolution process and deformation degree of the strain cloud map corresponding to the key abnormal section; determine a recovery strategy based on the evolution process of the strain cloud map and the deformation degree.

[0043] The above-mentioned tunnel structure detection method, device, computer equipment, readable storage medium and program product determine the key abnormal section in the tunnel structure; and based on a preset strain cloud map generation model and strain monitoring data corresponding to the key abnormal section, determine the first strain cloud map of the key abnormal section under the abnormal state, thereby achieving accurate positioning of the key abnormal section and extracting the strain deformation cloud map of the key section; and compare the first strain cloud map with the second strain cloud map under the normal state to obtain the strain cloud map evolution process and deformation degree corresponding to the key abnormal section; determine the recovery strategy based on the strain cloud map evolution process and deformation degree, and analyze the deformation evolution process of the tunnel structure, so as to take corresponding protective measures to ensure the operational safety of heavy-load railways and reduce the risk of sudden accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 1 is a flow chart of a tunnel structure detection method according to an embodiment;

[0046] Figure 2 1 is a flow chart of a tunnel structure detection method according to an embodiment;

[0047] Figure 3 1 is a flow chart of a tunnel structure detection method according to an embodiment;

[0048] Figure 4 A schematic diagram of the structure of a DIC-BOTDA device for heavy-load railway tunnel monitoring in one embodiment;

[0049] Figure 5 is a diagram showing the relationship between the DIC speckle strain cloud image and the BOTDA strain data in one embodiment;

[0050] Figure 6 is a structural block diagram of a tunnel structure detection device in one embodiment;

[0051] Figure 7 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] In an exemplary embodiment, as Figure 1 shown, a tunnel structure detection method is provided. In this embodiment, the method is exemplified by being applied to a terminal. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0054] Step 101: Determine the key abnormal section in the tunnel structure; and based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine the first strain cloud map of the key abnormal section in the abnormal state.

[0055] Among them, the tunnel structure may be the tunnel to be detected. The key abnormal section may refer to the section of the tunnel that may cause serious damage under abnormal disturbances. For example, the weak section of the tunnel structure, and the weak section may be the section with a higher surrounding rock grade and prone to uneven settlement. The preset strain cloud map generation model can reflect the spatial correlation between the strain monitoring data and the strain cloud map. The preset strain cloud map generation model may be a neural network model. For example, it may be a Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) joint neural network model. The strain monitoring data may be Brillouin optical time domain analysis strain monitoring data, and the strain monitoring data can reflect the deformation amount of the tunnel structure. The abnormal state may be that the tunnel structure is in an abnormal working condition. The first strain cloud map refers to the visualization image of the strain distribution of the tunnel structure in the abnormal state, and may be the cloud map displayed on the Digital Image Correlation (DIC) related device. The strain cloud map is a visualization graph that intuitively shows the strain distribution on the surface or inside of an object through color gradients. The first strain cloud map may be a DIC-BOTDA inversion strain cloud map.

[0056] Specifically, BOTDA distributed optical fibers can be arranged in the tunnel structure and 3D Digital Image Correlation (3D-DIC) cameras can be installed. The terminal can determine the key abnormal sections in the tunnel structure based on the BOTDA distributed optical fibers, and obtain the strain monitoring data of the BOTDA in the key abnormal sections. The terminal can input the strain monitoring data into a preset strain nephogram generation model to obtain the strain data vector corresponding to the key abnormal section, and perform surface fitting processing on the strain data vector to obtain the first strain nephogram.

[0057] Step 102: Compare the first strain nephogram with the second strain nephogram in the normal state to obtain the strain nephogram evolution process and deformation degree corresponding to the key abnormal section; determine the recovery strategy based on the strain nephogram evolution process and deformation degree.

[0058] Among them, the normal state is when the tunnel is in a healthy state. For example, it can be within the first year during the initial operation of the tunnel. The second strain nephogram is the strain distribution image when the tunnel structure is in a healthy state. The strain nephogram evolution process can be the change trend of the key abnormal section under abnormal disturbances. The deformation degree reflects the influence degree of the abnormal disturbance on the tunnel of the key abnormal section, and the deformation degree can include the damage type of the key abnormal section. The recovery strategy is the repair plan when the current abnormal disturbance affects the tunnel structure.

[0059] Specifically, the terminal can determine the strain nephogram dataset during the initial operation of the tunnel as the strain nephogram of the tunnel structure in the normal state. The terminal can obtain the second strain nephogram of the key abnormal section in the normal state from the strain nephogram dataset, and compare the first strain nephogram with the second strain nephogram to determine the strain nephogram evolution process and deformation degree corresponding to the key abnormal section. The terminal can determine the corresponding relationship between the strain nephogram evolution process, deformation degree and the recovery strategy. The terminal can determine the recovery strategy of the key abnormal section corresponding to the strain nephogram evolution process and deformation degree of the current key abnormal section in the corresponding relationship.

[0060] Optionally, the first strain nephogram and the second strain nephogram are in the same coordinate system.

[0061] The above tunnel structure detection method realizes the accurate positioning of the key abnormal section by determining the key abnormal section in the tunnel structure; and determining the first strain cloud map of the key abnormal section in the abnormal state based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, and can extract the strain deformation cloud map of the key section; and comparing the first strain cloud map with the second strain cloud map in the normal state to obtain the strain cloud map evolution process and deformation degree corresponding to the key abnormal section; determining the recovery strategy based on the strain cloud map evolution process and deformation degree, which can analyze the deformation evolution process of the tunnel structure, so as to take corresponding protection measures to ensure the operation safety of the heavy-haul railway and reduce the risk of sudden accidents.

[0062] In an exemplary embodiment, the tunnel structure detection method further includes:

[0063] Obtain the tunnel structure and determine the abnormal vibration signal characteristics of the tunnel structure; input the abnormal vibration signal characteristics into a preset vibration event classification model to determine the vibration event type corresponding to the abnormal vibration signal, and determine the risk level of the tunnel structure based on the vibration event type; when the risk level meets the preset abnormal conditions, execute the step of determining the key abnormal section in the tunnel structure.

[0064] Among them, the abnormal vibration signal characteristics are the feature vectors corresponding to the vibration signals in the tunnel structure when the vibration signals in the tunnel structure are abnormal vibration signals under abnormal interference. The preset vibration event classification model is a model for identifying the event type of abnormal interference vibration events. For example, it can be a backpropagation (BP) neural network classification model (Ad-BP) optimized by the Adaptive Moment Estimation (Adam) algorithm. The vibration event type is the type of abnormal working conditions. The risk level is the level of harm caused by the vibration event type to the tunnel structure. The preset abnormal conditions are used to characterize that the risk level corresponding to the vibration event type meets the conditions for causing damage to the tunnel structure.

[0065] Specifically, a distributed sensing optical cable is arranged in the tunnel structure. The distributed sensing optical cable can collect the vibration signals in the tunnel structure, analyze the vibration signals, determine the abnormal vibration signals in the vibration signals, extract the characteristics of the abnormal vibration signals, and obtain the abnormal vibration signal characteristics. The terminal can input the abnormal vibration signal characteristics into the preset vibration event classification model, output the probability values corresponding to multiple preset vibration event types respectively, and determine the vibration event type corresponding to the abnormal vibration signal characteristics as the preset vibration event type with the highest probability value. The terminal can pre-determine the corresponding relationship between the vibration event type and the risk level of the tunnel structure. In this way, the terminal can determine the current risk level of the current tunnel structure corresponding to the current vibration event type in the corresponding relationship.

[0066] If the risk level is greater than the preset risk threshold, it is determined that the risk level meets the preset abnormal condition, and the steps of the above embodiments are executed to determine the key abnormal section in the tunnel structure; and based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, the first strain cloud map of the key abnormal section in the abnormal state is determined. The first strain cloud map is compared with the second strain cloud map in the normal state to obtain the evolution process and deformation degree of the strain cloud map corresponding to the key abnormal section; a recovery strategy is determined based on the evolution process and deformation degree of the strain cloud map.

[0067] Optionally, the terminal can obtain the sample vibration signal characteristics and the corresponding sample vibration event types of the sample vibration signal characteristics. The terminal can input the sample vibration signal characteristics into the initial vibration event classification model to obtain the sample predicted vibration type, and optimize the parameters of the initial vibration event classification model based on the preset parameter optimization algorithm, the sample predicted vibration type, and the sample vibration type to obtain the preset vibration event classification model. For example, the preset parameter optimization algorithm can be the Adam optimization algorithm. The terminal can use the sample vibration signal characteristics and the corresponding sample vibration event types of the sample vibration signal characteristics as sample data, divide the sample data into a training set and a test set, train and learn the initial vibration event classification model based on the training set, and the test set is used to test the performance of the vibration event classification model.

[0068] In this embodiment, by determining the abnormal vibration signal and the vibration event type corresponding to the abnormal vibration signal, when the risk level of the vibration event meets the preset abnormal condition, the tunnel structure is analyzed for abnormalities, which improves the accuracy and efficiency of the tunnel structure abnormality investigation and further improves the safety of the tunnel structure.

[0069] In an exemplary embodiment, as Figure 2 shown, the specific implementation process of the step "obtain the tunnel structure and determine the abnormal vibration signal characteristics of the structure" may include:

[0070] Step 201, obtain the vibration signal of the tunnel structure; perform noise reduction processing on the vibration signal based on the preset spectral subtraction method to obtain the vibration signal after noise reduction processing; determine the endpoint detection threshold based on the vibration signal after noise reduction processing.

[0071] Among them, the vibration signal can be a Phase-sensitive Optical Time Domain Reflectometer ( -OTDR) vibration signal. The endpoint detection threshold can be used to determine whether the vibration signal is an abnormal vibration signal.

[0072] Specifically, the terminal can obtain vibration signals through a distributed sensing optical cable arranged in a tunnel structure. The terminal can perform Fourier transform processing on the vibration signals through a preset signal additive model of spectral subtraction to obtain the vibration signals after Fourier transform. The specific expression of the vibration signals after Fourier transform can be:

[0073]

[0074] Where, is the vibration signal after Fourier transform, is the effective signal after Fourier transform, is the noise signal after Fourier transform.

[0075] The terminal can perform frame addition and windowing processing on the vibration signals after Fourier transform to obtain the short-time Fourier transform result of the signals after frame addition and windowing processing. The specific expression of the short-time Fourier transform result of the signals after frame addition and windowing processing can be:

[0076]

[0077] Where, is the short-time Fourier transform result of the signals after frame addition and windowing processing, is the effective signal after frame addition and windowing processing, is the noise signal after frame addition and windowing processing.

[0078] Square both sides of the specific expression of the short-time Fourier transform result of the signals after frame addition and windowing processing, and the relationship of the short-time power spectra of the three signals is:

[0079]

[0080] Where, represents the square modulus of the signal after frame addition and windowing processing, is the square of the modulus of the effective signal after frame addition and windowing processing, is the square of the modulus of the noise signal after frame addition and windowing processing.

[0081] Since the effective signal and the noise signal are independent of each other, so there is: Therefore, the specific expression of the short-time power spectrum of the effective signal can be:

[0082]

[0083] It is known that the number of frames of the undisturbed signal is N, then the mean value of the short-time power spectrum of the noise signal is represented by The mean value of the short-time power spectrum of the noise signal The specific expression can be:

[0084]

[0085] Therefore, the specific expression for determining the short-time power spectrum of the effective signal can be:

[0086]

[0087] Where α and β are the over-subtraction factor and the spectral lower limit factor respectively, SNR is the signal-to-noise ratio, α0 is the value when SNR = 0, and 1 / r is the change rate.

[0088] The terminal can determine the effective signal as the vibration signal after noise reduction processing, and determine the short-time power spectrum of the effective signal as the short-time power spectrum of the vibration signal after noise reduction processing. Based on the short-time power spectrum and the periodic signal of the vibration signal, determine the power spectral density of the vibration signal after noise reduction processing. The specific calculation formula for the power spectral density can be:

[0089]

[0090] Where PSD is the power spectral density and T is the signal period.

[0091] The terminal can calculate the root mean square value PSD-RMS of the power spectral density PSD, calculate the statistical values of all PSD-RMS values. The statistical values can include the average value, median value, maximum value, minimum value, standard deviation, etc. The terminal can determine the fluctuation range of the PSD-RMS value based on the statistical values, and determine the median of the fluctuation range as the endpoint detection threshold. Optionally, the terminal can determine the fluctuation range of the PSD-RMS value by means of a confidence interval and in combination with the statistical values.

[0092] Step 202: Perform abnormal screening processing on the vibration signal after noise reduction processing based on the endpoint detection threshold to obtain abnormal vibration signals.

[0093] Specifically, the terminal can perform signal interception on the vibration signal after noise reduction processing based on the endpoint detection threshold to obtain multiple vibration signal segments, determine the vibration signal segments greater than or equal to the endpoint detection threshold as abnormal vibration signals, and determine the vibration signal segments less than the endpoint detection threshold as normal vibration signals.

[0094] Step 203: Perform feature extraction on the abnormal vibration signals based on the local characteristic scale decomposition method combined with multi-scale permutation entropy to obtain abnormal vibration signal features.

[0095] Specifically, the terminal can use the Local Characteristic Scale Decomposition (LCD) method to decompose the abnormal vibration signal into multiple Intrinsic Scale Components (ISC) and a residual. The specific expression of the abnormal vibration signal can be:

[0096]

[0097] where ψ(μ) is the abnormal vibration signal, is the intrinsic scale component, η is the number of intrinsic scale components, c ρ is the ρ-th intrinsic scale component, and r η (μ) is the residual.

[0098] Each intrinsic scale component satisfies the following conditions: (1) The product of all adjacent two extreme points of the abnormal vibration signal is negative; (2) Assuming that the abnormal vibration signal has M extreme points, represented by (τ ρ , ψ ρ )(ρ = 1, 2... M), where τρ is the time of the ρ-th extreme point, and ψ ρ is the amplitude of the ρ-th extreme point. Determine a linear function from (τ ρ , ψ ρ ) and (τ ρ+2 , ψ ρ+2 ), and obtain the function value A k+1 at τ k+1 . Ensure that the ratio of all A ρ+1 to the extreme values corresponding to τ ρ+1 is consistent. The specific expression of the function value A ρ+1 can be:

[0099]

[0100] According to the definition of ISC, the specific process of decomposition based on LCD can include the following steps:

[0101] The terminal determines all the extreme points ψ of the abnormal vibration signal ψ(μ) and the corresponding moments of the extreme points τ ρ , solves the function value A ρ+1 corresponding to the extreme point with the opposite polarity between any two extreme points with the same polarity of ψ(μ), and then obtains the function value L ρ+1 of the baseline signal at this moment according to a specific proportional relationship. The specific expression of the function value L ρ+1 of the baseline signal at this moment can be:

[0102]

[0103] Among them, the parameter generally takes a value of 0.5, and ψ ρ+1 is the (ρ + 1)-th extreme value.

[0104] Calculate the first difference between the function value of the baseline signal at this moment and the function value of the baseline signal at the previous moment, calculate the second difference between the extreme point of the abnormal vibration signal at this moment and the extreme point of the abnormal vibration signal at the previous moment, calculate the third difference between x μ and the extreme point of the abnormal vibration signal at the previous moment, calculate the ratio of the first difference to the second difference, calculate the product value of this ratio and the third difference, and determine the sum value of this product value and the function value of the baseline signal at the previous moment as the baseline segment signal. The specific expression of the baseline segment signal H ρ can be:

[0105]

[0106] Among them, x μ is the original signal vector, and μ is the original signal.

[0107] Connect H ρ to obtain H1(μ), and separate H1(μ) from the abnormal vibration signal ψ(μ) to obtain the residual signal T1(μ). The specific expression of the residual signal can be:

[0108] Τ1(μ) = ψ(μ) - H1(μ)

[0109] If T1(μ) meets the ISC component discrimination condition, then make c1(μ) = T1(μ); if it does not meet, then continue to loop T1(μ) according to the above steps until the g-th T 1g (μ), T 1g (μ) represents the first ISC component c1(μ) of the signal T1(μ).

[0110] Separate c1(μ) from the abnormal vibration signal ψ(μ) to obtain the updated abnormal vibration signal γ1(μ). The specific expression of γ1(μ) can be:

[0111] Υ1(μ) = ψ(μ) - c1(μ)

[0112] Loop the above all steps in this embodiment j times until the number of extreme points of the γ 1g (μ) at the g-th time is 0 or the standard deviation SD ≤ 0.3 for all points, then stop the loop to obtain the final abnormal vibration signal. The specific formula for calculating the standard deviation can be:

[0113]

[0114] Among them, U is the number of all points, T 1g (μ) The residual signal at the g-th time, T 1(g-1) (μ) The residual signal at the (g - 1)-th time.

[0115] Multiscale Permutation Entropy (MPE) is used to extract features from the updated abnormal vibration signal to obtain the abnormal vibration signal features. Assuming the current updated abnormal vibration signal is W = (w(r), r = 1, 2, 3, …, d), the specific process of feature extraction can include the following steps:

[0116] Perform the coarse-graining principle processing on the updated abnormal vibration signal W to obtain the signal sequences under different scale factors. The specific calculation formula of the coarse-graining principle can be:

[0117]

[0118] Among them, φ s (q) is the signal sequence under different scale factors, S is the scale factor, and q is the signal symbol sequence.

[0119] Perform time reconstruction on the signal sequences under different scale factors to obtain the time-reconstructed signal sequences. The specific formula for time reconstruction can be:

[0120]

[0121] Among them, is the time-reconstructed signal sequence, m is the embedding dimension, and τ is the delay time.

[0122] Calculate the Error Percentage (PE) values of each time-reconstructed signal sequence. The specific formula for calculating the PE value can be:

[0123]

[0124] Among them, is the PE value of each time-reconstructed signal sequence, m is the embedding dimension, is the probability of the q-th symbol sequence appearing.

[0125] Perform normalization processing on the calculated PE values to obtain the normalized PE values. The specific formula for normalization processing can be:

[0126]

[0127] Among them, is the normalized PE value, The larger the value, the more random the signal, indicating that the signal is more complex.

[0128] According to Calculate the Maximum Permissible Error (MPE) value of the vibration signal based on the value. Obtain the MPE value of the parameter corresponding to the vibration characteristic of the signal, convert the MPE value into the threshold value of the corresponding vibration parameter, and then identify the anomaly through feature extraction and comparison. Identify the vibration characteristics under the tunnel operation conditions through time-frequency domain analysis, and verify whether its energy exceeds the allowable range mapped by the MPE. If it exceeds the MPE threshold, it is considered that abnormal vibration has occurred, and the terminal can construct the abnormal vibration signal feature with the MPE value under the abnormal vibration.

[0129] In this embodiment, by screening out the abnormal vibration signals from the vibration signals in the tunnel structure and extracting the signal features of the abnormal vibration signals, the abnormal vibration signals are obtained, so as to determine the influence of the abnormal disturbance on the tunnel structure and determine the corresponding recovery strategy.

[0130] In an exemplary embodiment, the specific implementation process of step 101, "Based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine the first strain cloud map of the key abnormal section in the abnormal state" may include:

[0131] Based on the preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine the strain data vector corresponding to the key abnormal section; restore the strain data vector to the three-dimensional space where the strain monitoring data is located to obtain the target strain data vector, and perform surface fitting processing on the target strain data vector based on the preset surface fitting algorithm to obtain a non-uniform surface; perform reconstruction processing on the non-uniform surface based on the surface network segmentation energy algorithm to obtain the first strain cloud map of the key abnormal section in the abnormal state.

[0132] Among them, the preset surface fitting algorithm at least includes a first-order derivative vector continuous algorithm. The non-uniform surface can be a Non-Uniform Rational B-Splines (NURBS) surface.

[0133] Specifically, the terminal can input the strain monitoring data corresponding to the key abnormal section into the preset strain cloud map data generation model, and output the strain data vector corresponding to the key abnormal section. The terminal can perform three-dimensional reconstruction on the strain data vector and restore the strain data vector to the three-dimensional cloud map where the strain monitoring data is located to obtain the target strain data vector.

[0134] The terminal can pre-define the shape value points of the curve as B χ, χ = 0, 1, …, v, where χ is the number of curve fittings. The profile points can be key points used to determine the curve shape. According to the curve fitting algorithm, the target strain data vector is fitted into a NURBS curve through the profile points. The specific expression of the NURBS curve can be:

[0135]

[0136] where C(u) is the NURBS curve, P i represents the i-th target strain data vector; v represents the curve order; M i,t (u) is the t-th normalized B-spline basis function; U is the knot vector.

[0137] The terminal calculates the parameter value U k corresponding to the property point C based on the chord length parameterization method, and calculates the profile point B k of the curve by the average value method. The calculation formula can be: χ

[0138]

[0139] Given that the coordinates of the target strain data vector P i are three-dimensional structures, the calculation formula of the profile point B χ is simultaneously combined with the following two linear equations to find the target strain data vector P i i .

[0140]

[0141] Based on the calculated P i and the calculation formula of the profile points, the NURBS curve is determined. The expression formula of this curve can be {B χ，l}, χ = 0, 1, … v, l = 0, 1, … c, where l is the l-th curve of the surface and c is the c-th profile point of the l-th curve. Based on the NURBS curve interpolation, a non-uniform rational B-spline surface is formed: The specific calculation formula of the NURBS surface can be:

[0142]

[0143] In the formula, S is the surface; u k , v l are parameter values; B χ,l is the NURBS curve.

[0144] The solution of the NURBS surface control vertices usually first interpolates the point cloud data {B χ，l}, χ = 0, 1, … v, l = 0, … c along the U direction to find the control point cloud {R i,l}.

[0145]

[0146] Where M i,t (u) is the t-order canonical B-spline basis function; M j,q (π) is the q-order canonical B-spline basis function.

[0147] The terminal can reconstruct the non-uniform surface based on the surface network sharding energy algorithm to obtain the first strain cloud map of the key abnormal section under the abnormal state.

[0148] In an exemplary embodiment, the specific implementation process of the step of "reconstructing the non-uniform surface based on the surface network slicing energy algorithm to obtain the first strain cloud map of the key abnormal section under the abnormal state" may include:

[0149] An original surface is constructed based on the target strain data vector. The non-uniform surface is constructed based on local bicubic surface interpolation to obtain multiple bicubic Bezier surface patches, which are then spliced to obtain an auxiliary surface. The auxiliary surface is roughly fitted based on a preset smoothing threshold to obtain smooth and non-smooth areas. Grid lines are constructed in the non-smooth areas, and the bicubic Bezier surface patches corresponding to the non-smooth areas are removed. The grid lines corresponding to the non-smooth areas are locally smoothed to obtain target grid lines. The target grid lines are adjusted based on the normal distance between the auxiliary surface and the original surface to obtain adjusted target grid lines. The grid lines in the first target direction of the adjusted target grid lines are determined as the main curve family, and the grid lines in the second target direction of the adjusted target grid lines are determined as the guide curve family. The auxiliary surface is optimized based on the main curve family and the guide curve family to obtain a target auxiliary surface, and the boundary of the target auxiliary surface meets the first-order continuity condition with the boundary of the original surface. The target strain data vector and the target auxiliary surface are fitted to obtain a first strain contour.

[0150] Specifically, the terminal can construct the original surface based on the target strain data vector, perform surface construction processing on the non-uniform surface based on local bicubic surface interpolation, obtain multiple bicubic Bezier surface patches, and splice the multiple bicubic Bezier surface patches to obtain the auxiliary surface for the data series of the given type value point. The specific calculation formula of the auxiliary surface can be:

[0151]

[0152] Among them, P i,j is the target strain data vector. For auxiliary surfaces. is the parameter value, c and v are the control vertices of the node vector node line, Ni,3 , N j,3 is the basis function.

[0153] The terminal can roughly fit the auxiliary surface based on a preset fairing threshold to obtain a fairing region and a non-fairing region; construct grid lines in the non-fairing region, remove the bicubic B-spline surface patches corresponding to the non-fairing region, and locally fair the grid lines corresponding to the non-fairing region to obtain target grid lines; adjust the target grid lines based on the normal distance between the auxiliary surface and the original surface to obtain the adjusted target grid lines; determine the main curve family for the grid lines in the first target direction of the adjusted target grid lines, and determine the guiding curve family for the grid lines in the second target direction of the adjusted target grid lines; optimize the auxiliary surface based on the main curve family and the guiding curve family to obtain a target auxiliary surface, where the boundary of the target auxiliary surface conforms to the first-order continuity condition with the boundary of the original surface; fit the target strain data vector and the target auxiliary surface to obtain the first strain nephogram.

[0154] In this embodiment, converting the non-uniform surface into a strain nephogram realizes the visualization of the strain nephogram.

[0155] In an exemplary embodiment, as Figure 3 shown, the tunnel structure detection method further includes the following steps:

[0156] Step 301, determine the sample section of the tunnel structure, and obtain the initial sample strain nephogram data and the initial sample strain monitoring data under the normal state.

[0157] Among them, the sample section can be a weak section of the tunnel structure, and this weak section can be a section with a higher surrounding rock grade and prone to uneven settlement, etc. The normal state can be the state of the tunnel during the initial operation period.

[0158] Specifically, during the operation of the tunnel structure, BOTDA distributed optical fibers are laid in the tunnel structure in a stacked manner, and speckles are sprayed on the surface concrete of the BOTDA distributed optical fiber monitoring points in the sample area, as Figure 4 shown. Determine the number of detection points to be selected according to the size of the BOTDA spatial resolution and the size of the speckles. The speckle area can contain Z BOTDA monitoring points, and the area of these Z features is used as the key (Region Of Interest, ROI area). A 3D-DIC camera is fixedly placed on the construction fixed platform above the tunnel for a long time to take pictures and scan the speckle area. Optionally, the ROI area is adjusted according to the position of the diseased sensor in the BOTDA distributed optical fiber. As Figure 5 shown, Figure 5The x-axis, y-axis, and z-axis are the coordinate axes where the tunnel structure is located. The square is the three-dimensional cloud map of the DIC tunnel surface, and the range of this three-dimensional area is 0.1m × 0.1m. The circles in the figure are the strain monitoring points (Z BOTDA monitoring points).

[0159] The terminal can obtain the initial strain cloud map of the speckle area through a 3D-DIC camera and simultaneously extract the initial strain monitoring data monitored by BOTDA distributed optical fiber at this time point. Optionally, taking the initial year of tunnel operation as the healthy state, about Z × 11000 initial strain monitoring data of BOTDA and a large number of DIC three-dimensional strain cloud maps are obtained, and these data are saved and a data set is formed.

[0160] Define the matrix of the initial strain monitoring data set as Y. If the number of sampling time points is n and the number of strain sensors is X, then the long-term initial strain monitoring data set Y of this sample section of the tunnel belongs to R X×n , and the expression of the initial strain monitoring data set can be:

[0161] Y = [ε 1 , ε 2 , … ε o , ], o ∈ (1, 2, …, X)

[0162] where o is any measurement point of the strain of the tunnel BOTDA structure; ε o is the vector composed of the long-term monitored initial strain monitoring data of any strain measurement point o. ε i is the vector composed of the initial strain monitoring data of the i-th monitoring measurement point, k ∈ (1, 2, …, n), where k is any monitoring time point of the strain of the tunnel structure. is the initial strain detection data of any strain measurement point o at the k-th monitoring time point.

[0163] Step 302: Normalize the initial sample strain cloud map data corresponding to the sample section to obtain the sample strain cloud map data; standardize the initial sample strain monitoring data in the sample section to obtain the target sample strain monitoring data.

[0164] Specifically, the terminal can normalize the pixel values in the initial sample strain cloud map data to obtain the normalized initial sample strain cloud map, and at the same time normalize the data labels of the normalized initial sample strain cloud map to obtain the sample strain cloud map data. Optionally, the data label can be the position of the pixel in the cloud map image.

[0165] The terminal performs standardized processing on the initial strain monitoring data to obtain the initial strain monitoring data after standardized processing, and it can be determined that the initial strain monitoring data after standardized processing is the target sample strain monitoring data. The specific formula for the target sample strain monitoring data can be:

[0166]

[0167] Among them, is the BOTDA strain monitoring data vector of any strain measurement point i at the k-th monitoring time point after data standardized processing, is the mean value of the monitoring vector of any strain measurement point o; is the variance of the monitoring vector of any strain measurement point o.

[0168] Optionally, the pixel values in the initial sample strain nephogram data are processed from the [0, 255] region to the [0, 1] interval, and the data labels (X, Y) are normalized, and the strain values are adjusted to the [0, 1] interval.

[0169] Step 303: Based on a preset conversion algorithm, perform coordinate conversion on the sample strain nephogram data to obtain the target sample strain nephogram data.

[0170] Among them, the preset conversion algorithm can be the two-dimensional image reference (Image Reference 2D, imref2d) function.

[0171] Specifically, the terminal can perform coordinate conversion on the sample strain nephogram data based on the imref2d function to obtain the target sample strain nephogram data.

[0172] Optionally, the terminal can save the target sample strain nephogram data and the target sample strain monitoring data at the corresponding position as sample data to the database.

[0173] Step 304: Input the target sample strain monitoring data into the initial strain nephogram generation model to obtain the predicted strain nephogram data, and correct the initial strain nephogram generation model based on the predicted strain nephogram data and the target sample strain nephogram data to obtain the preset strain nephogram generation model.

[0174] Specifically, the terminal inputs the target sample strain monitoring data into the initial strain nephogram generation model to obtain the predicted strain nephogram data, and corrects the initial strain nephogram generation model based on the predicted strain nephogram data and the target sample strain nephogram data to obtain the preset strain nephogram generation model. In addition, during the training process of the initial strain generation model, the parameters of the initial strain generation model can be continuously adjusted through the backpropagation algorithm. The rectified linear unit (ReLU) function can be used as the activation function during the training process.

[0175] Optionally, since there is a certain spatial correlation between the strain monitored by BOTDA and the surface strain of concrete, the specific expression of the strain vector ω of the BOTDA surface concrete can be:

[0176]

[0177] where G is the transfer function between the surface concrete strain data and the BOTDA strain data, is the BOTDA strain monitoring data vector.

[0178] In the 3D-DIC strain nephogram data vector, ω is part of the data vector. Therefore, there is a certain correlation between this vector and the 3D-DIC strain nephogram data vector ξ, and the specific expression can be:

[0179] ξ = U(ω)

[0180] where ξ is the 3D-DIC strain nephogram data vector and U is the correlation function of the strain detection data in the same area.

[0181] Based on this, the expression of the relationship between the 3D-DIC strain nephogram data vector and the strain vector of the BOTDA surface concrete can be:

[0182]

[0183] where L(·) is the transfer function between the BOTDA strain data and the 3D-DIC strain nephogram data. Therefore, there is a certain spatial correlation between the 3D-DIC strain nephogram data vector and the BOTDA strain data.

[0184] In one embodiment, it is possible to extract and analyze the strain nephogram of the structural deformation of the key section caused by the illegal construction operations within the protected area of the heavy-haul railway tunnel, so as to achieve the purpose of monitoring the illegal construction operations in the tunnel protected area.

[0185] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are sequentially shown according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps or stages in other steps.

[0186] Based on the same inventive concept, an embodiment of the present application further provides a tunnel structure detection device for implementing the tunnel structure detection method involved above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the tunnel structure detection device provided below can refer to the limitations on the tunnel structure detection method in the above text and will not be elaborated here.

[0187] In an exemplary embodiment, as Figure 6 shown, a tunnel structure detection device 60 is provided, including: a determination module 61 and a detection module 62, where:

[0188] The determination module 61 is configured to determine the key abnormal section in the tunnel structure; and based on a preset strain cloud map generation model and the strain monitoring data corresponding to the key abnormal section, determine the first strain cloud map of the key abnormal section in the abnormal state;

[0189] The detection module 62 is configured to compare the first strain cloud map with the second strain cloud map in the normal state to obtain the strain cloud map evolution process and the deformation degree corresponding to the key abnormal section; and determine a recovery strategy based on the strain cloud map evolution process and the deformation degree.

[0190] In one of the embodiments, the determination module 61 is further configured to obtain the tunnel structure and determine the abnormal vibration signal characteristics of the tunnel structure;

[0191] Input the abnormal vibration signal characteristics into a preset vibration event classification model to determine the vibration event type corresponding to the abnormal vibration signal, and based on the vibration event type, determine the risk level of the tunnel structure;

[0192] When the risk level meets the preset abnormal conditions, execute the step of determining the key abnormal section in the tunnel structure.

[0193] In one of the embodiments, the determination module 61 is further configured to obtain the vibration signal of the tunnel structure; perform noise reduction processing on the vibration signal based on a preset spectral subtraction method to obtain the vibration signal after noise reduction processing; determine an endpoint detection threshold based on the vibration signal after noise reduction processing;

[0194] Perform abnormal screening processing on the vibration signal after noise reduction processing based on the endpoint detection threshold to obtain an abnormal vibration signal;

[0195] Extract features from the abnormal vibration signal based on the local characteristic scale decomposition method combined with multi-scale permutation entropy to obtain abnormal vibration signal characteristics.

[0196] In one embodiment, a determination module 61 is configured to determine a strain data vector corresponding to the key abnormal section based on a preset strain nephogram generation model and strain monitoring data corresponding to the key abnormal section;

[0197] Restore the strain data vector to the three-dimensional space where the strain monitoring data is located to obtain a target strain data vector, and perform surface fitting processing on the target strain data vector based on a preset surface fitting algorithm to obtain a non-uniform surface;

[0198] Perform reconstruction processing on the non-uniform surface based on a surface network segmentation energy algorithm to obtain a first strain nephogram of the key abnormal section in an abnormal state.

[0199] In one embodiment, the determination module 61 is configured to construct an original surface based on the target strain data vector, perform surface construction processing on the non-uniform surface based on local bicubic surface interpolation to obtain a plurality of bicubic Bezier surface patches, and splice the plurality of bicubic Bezier surface patches to obtain an auxiliary surface;

[0200] Perform rough fitting on the auxiliary surface based on a preset fairing threshold to obtain a fairing area and an unfairing area;

[0201] Construct grid lines in the unfairing area, remove the bicubic Bezier surface patches corresponding to the unfairing area, and perform local fairing processing on the grid lines corresponding to the unfairing area to obtain target grid lines;

[0202] Adjust the target grid lines based on the normal distance between the auxiliary surface and the original surface to obtain adjusted target grid lines;

[0203] Determine the grid lines in the first target direction of the adjusted target grid lines as the main curve family, and determine the grid lines in the second target direction of the adjusted target grid lines as the guiding curve family; optimize the auxiliary surface based on the main curve family and the guiding curve family to obtain a target auxiliary surface, and the boundary of the target auxiliary surface conforms to the first-order continuity condition with the boundary of the original surface;

[0204] Fit the target strain data vector and the target auxiliary surface to obtain a first strain nephogram.

[0205] In one embodiment, the determination module 61 is further configured to determine a sample section of the tunnel structure, and obtain initial sample strain nephogram data and initial sample strain monitoring data in a normal state;

[0206] Normalize the initial sample strain nephogram data corresponding to the sample section to obtain sample strain nephogram data; standardize the initial sample strain monitoring data in the sample section to obtain target sample strain monitoring data;

[0207] Based on a preset conversion algorithm, perform coordinate conversion on the sample strain nephogram data to obtain target sample strain nephogram data;

[0208] Input the target sample strain monitoring data into the initial strain nephogram generation model to obtain predicted strain nephogram data, and correct the initial strain nephogram generation model based on the predicted strain nephogram data and the target sample strain nephogram data to obtain a preset strain nephogram generation model.

[0209] Each module in the above tunnel structure detection device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0210] In an exemplary embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 7 shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, a mobile cellular network, near field communication (NFC), or other technologies. The computer program, when executed by the processor, implements a tunnel structure detection method. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the outer shell of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0211] Those skilled in the art can understand that Figure 7 The structure shown in Figure 7 is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0212] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0213] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0214] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0215] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0216] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0217] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope recorded in the present application.

[0218] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A tunnel structure detection method, characterized in that, The method includes: Determining a key abnormal section in the tunnel structure; and determining a first strain contour map of the key abnormal section in the abnormal state based on a preset strain contour map generation model and the strain monitoring data corresponding to the key abnormal section; Comparing the first strain contour map with a second strain contour map in the normal state to obtain the evolution process and deformation degree of the strain contour map corresponding to the key abnormal section; and determining a recovery strategy based on the evolution process of the strain contour map and the deformation degree.

2. The method according to claim 1, characterized in that, The method further includes: Obtaining the tunnel structure and determining the abnormal vibration signal characteristics of the tunnel structure; Inputting the abnormal vibration signal characteristics into a preset vibration event classification model to determine the type of vibration event corresponding to the abnormal vibration signal, and determining the risk level of the tunnel structure based on the type of vibration event; When the risk level meets a preset abnormal condition, performing the step of determining the key abnormal section in the tunnel structure.

3. The method according to claim 2, wherein The obtaining of the tunnel structure and the determination of the abnormal vibration signal characteristics of the structure include: Obtaining the vibration signal of the tunnel structure; performing noise reduction processing on the vibration signal based on a preset spectral subtraction method to obtain the vibration signal after noise reduction processing; and determining an endpoint detection threshold based on the vibration signal after noise reduction processing; Performing abnormal screening processing on the vibration signal after noise reduction processing based on the endpoint detection threshold to obtain an abnormal vibration signal; Performing feature extraction on the abnormal vibration signal based on the local characteristic scale decomposition method combined with multi-scale permutation entropy to obtain abnormal vibration signal characteristics.

4. The method according to claim 1, wherein The determining of the first strain contour map of the key abnormal section in the abnormal state based on a preset strain contour map generation model and the strain monitoring data corresponding to the key abnormal section includes: Determining a strain data vector corresponding to the key abnormal section based on a preset strain contour map generation model and the strain monitoring data corresponding to the key abnormal section; Restoring the strain data vector to the three-dimensional space where the strain monitoring data is located to obtain a target strain data vector, and performing surface fitting processing on the target strain data vector based on a preset surface fitting algorithm to obtain a non-uniform surface; Performing reconstruction processing on the non-uniform surface based on a surface network segmentation energy algorithm to obtain the first strain contour map of the key abnormal section in the abnormal state.

5. The method according to claim 4, wherein The performing of reconstruction processing on the non-uniform surface based on a surface network segmentation energy algorithm to obtain the first strain contour map of the key abnormal section in the abnormal state includes: Constructing an original surface based on the target strain data vector, performing surface construction processing on the non-uniform surface based on local bicubic surface interpolation to obtain a plurality of bicubic B-spline surface patches, and splicing the plurality of bicubic B-spline surface patches to obtain an auxiliary surface; Performing rough fitting on the auxiliary surface based on a preset smoothing threshold to obtain a smooth region and a non-smooth region; Constructing grid lines in the non-smooth region, removing the bicubic B-spline surface patches corresponding to the non-smooth region, and performing local smoothing processing on the grid lines corresponding to the non-smooth region to obtain target grid lines; Adjust the target grid line based on the normal distance between the auxiliary surface and the original surface to obtain the adjusted target grid line; Determine the grid lines in the first target direction of the adjusted target grid line as the main curve family, and determine the grid lines in the second target direction of the adjusted target grid line as the guiding curve family; optimize the auxiliary surface based on the main curve family and the guiding curve family to obtain a target auxiliary surface, and the boundary of the target auxiliary surface conforms to the first-order continuity condition with the boundary of the original surface; Fit the target strain data vector and the target auxiliary surface to obtain a first strain nephogram.

6. The method according to claim 1, wherein The method further includes: Determine a sample section of the tunnel structure, and obtain initial sample strain nephogram data and initial sample strain monitoring data under normal conditions; Normalize the initial sample strain nephogram data corresponding to the sample section to obtain sample strain nephogram data; standardize the initial sample strain monitoring data in the sample section to obtain target sample strain monitoring data; Perform coordinate transformation on the sample strain nephogram data based on a preset transformation algorithm to obtain target sample strain nephogram data; Input the target sample strain monitoring data into an initial strain nephogram generation model to obtain predicted strain nephogram data, and correct the initial strain nephogram generation model based on the predicted strain nephogram data and the target sample strain nephogram data to obtain a preset strain nephogram generation model.

7. A tunnel structure detection device, characterized in that, The device includes: A determination module, configured to determine a key abnormal section in the tunnel structure; and determine a first strain nephogram of the key abnormal section in an abnormal state based on a preset strain nephogram generation model and the strain monitoring data corresponding to the key abnormal section; A detection module, configured to compare the first strain nephogram with a second strain nephogram in a normal state to obtain the strain nephogram evolution process and the deformation degree corresponding to the key abnormal section; determine a recovery strategy based on the strain nephogram evolution process and the deformation degree.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. 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 according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images