Weak surrounding rock tunnel initial support cavity detection method and system

Abstract

The application belongs to the technical field of tunnel engineering detection, and specifically discloses a soft surrounding rock tunnel initial support cavity detection method and system, which comprises the following steps: collecting data and calibrating positions through a radar detection vehicle, analyzing local electromagnetic wave characteristics to determine cavities and mark suspicious areas, extracting time differences to calculate cavity depths in combination with electromagnetic wave speeds, determining longitudinal lengths and transverse widths according to radar images, calculating initial detection volumes in combination with a cavity shape pattern library, confirming actual parameters through secondary reexamination, finally evaluating structure risk levels according to actual positions and volumes, and generating a detection report; the radar detection vehicle is used for non-contact geological radar detection, so that physical damage of a support structure caused by traditional drilling is avoided, the integrity of the support structure is ensured, and through quantitative criteria of multiple characteristics such as reflected wave amplitude, phase and waveform, subjective artificial judgment is replaced, the objective reliability of detection results is improved, and the detection efficiency is greatly improved in the automatic and continuous scanning mode.

Description

Technical Field

[0001] This invention belongs to the field of tunnel engineering detection technology, and relates to a method and system for detecting cavities in the initial support of tunnels with weak surrounding rock. Background Technology

[0002] Initial support for tunnels in weak surrounding rock refers to a crucial support system implemented immediately after excavation in tunnels with poor geological conditions and weak self-stabilizing capacity. This system aims to control rock deformation and prevent collapse and instability. It forms the main load-bearing structure during tunnel construction by co-deforming and sharing loads with the surrounding rock. Due to the low strength and large deformation of weak surrounding rock, the structural integrity of the initial support directly affects construction safety and long-term stability. If cavities exist behind the support structure, their support effect will be significantly weakened, potentially leading to localized damage or a chain reaction of collapses. Therefore, cavity detection is necessary to mitigate these risks.

[0003] For example, Chinese invention patent CN113219052B discloses a method for identifying cavities in the initial support of tunnels in weak surrounding rock. This method targets the accessible area of ​​the initial support, using manual tapping to detect hollow sounds to locate suspicious areas. These suspicious areas are then verified by drilling, and cavities are confirmed based on whether loose cuttings fall from the boreholes. For inaccessible areas, discrete drilling points are arranged at intervals of less than 100 centimeters. Cavities are initially identified by sudden changes in drill bit speed and the phenomenon of loose cuttings falling. If necessary, a second hole is drilled 10-15 centimeters above the first borehole to further confirm the location of the cavity.

[0004] The existing technologies mentioned above have the following shortcomings: 1. Currently, they rely on manual tapping and intensive drilling operations. The drilling process can damage the physical structure of the initial support, which can easily induce local cracking of the support or secondary deformation of the surrounding rock, thus exacerbating the construction safety risks. Secondly, the judgment of hollow sound from manual tapping and the perception of sudden changes in drilling speed during drilling both depend on the experience of the inspectors and lack a unified and quantitative judgment standard. This results in highly subjective and unreliable test results. At the same time, the combined mode of discrete drilling and manual operation has extremely low detection efficiency and is difficult to meet the rapid detection needs of long-distance soft surrounding rock tunnels.

[0005] 2. Currently, obtaining cavity information through discrete borehole points can only achieve qualitative identification of the presence or absence of isolated points, and cannot form continuous spatial distribution data of cavities. It is very easy to miss small-sized cavities or cavity edge areas due to the spacing between boreholes, resulting in a high rate of missed detection. At the same time, it can only locate the approximate location of cavities, and cannot accurately measure key quantitative parameters such as the depth, longitudinal extension length, lateral width and volume of cavities. Consequently, it is impossible to carry out targeted safety risk assessments, making it difficult to guide the accurate formulation of subsequent cavity remediation plans. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background technology, a method and system for detecting cavities in the initial support of tunnels in weak surrounding rock is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: This invention provides a method for detecting cavities in the initial support of tunnels in weak surrounding rock, including: S1, calibrating the position of tunnel contour point cloud data collected by a radar detection vehicle and standard tunnel contour data, and obtaining ground-penetrating radar detection data.

[0008] S2. Perform imaging analysis on the local electromagnetic wave characteristics in the ground-penetrating radar detection data to determine whether there are cavities. When a cavity is determined to exist, mark the corresponding location as a suspected cavity area and record the initial detection data.

[0009] S3. For the suspected cavity area, the depth of the cavity is determined based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave velocity, and the longitudinal extension length and lateral width of the cavity are determined based on the radar image.

[0010] S4. Calculate the initial volume based on the depth, longitudinal extension length, and lateral width of the cavity, combined with a preset cavity shape graphic library.

[0011] S5. Conduct a second inspection of the suspected cavity area, compare and analyze the consistency between the second inspection data and the initial inspection data, determine the actual location coordinates and actual volume of the cavity, assess the structural risk level accordingly, and generate a cavity detection report.

[0012] The present invention also provides a cavity detection system for the initial support of tunnels in weak surrounding rock, comprising: a point cloud calibration module, which performs position calibration based on the tunnel contour point cloud data collected by the radar detection vehicle and the tunnel contour standard data, and acquires ground-penetrating radar detection data.

[0013] The cavity identification module performs imaging analysis on the local electromagnetic wave features in the ground-penetrating radar detection data to determine whether cavities exist. When a cavity is determined to exist, the corresponding location is marked as a suspected cavity area, and the initial detection data is recorded.

[0014] The parameter extraction module determines the depth of the suspected cavity based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave velocity, and determines the longitudinal extension length and lateral width of the cavity based on the radar image.

[0015] The volume calculation module calculates the initial volume based on the depth, vertical extension length, and horizontal width of the cavity, combined with a preset cavity shape graphic library.

[0016] The re-inspection report module performs a second re-inspection on the suspected cavity area, compares and analyzes the consistency between the re-inspection data and the initial inspection data, determines the actual location coordinates and actual volume of the cavity, assesses the structural risk level accordingly, and generates a cavity detection report.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention uses a non-contact geological radar detection vehicle to avoid physical damage to the support structure caused by traditional drilling, thus ensuring its integrity. At the same time, the quantitative criteria of multi-feature fusion such as reflected wave amplitude, phase and waveform replace subjective manual judgment, thereby improving the objective reliability of the detection results. The automated continuous scanning mode greatly improves the detection efficiency.

[0018] (2) This invention achieves precise positioning of the radar detection vehicle by comparing single points with the overall contour based on tunnel contour point cloud and standard data, calculating pose deviation and performing coordinate correction, ensuring accurate correspondence between radar detection data and actual tunnel spatial position, providing a reliable spatial reference for subsequent parameter calculation, and fundamentally improving detection accuracy.

[0019] (3) The present invention determines the cavity depth, longitudinal extension length and transverse width item by item through a precise algorithm, and calculates the volume based on the three-dimensional envelope and intelligent matching strategy, thereby obtaining continuous and accurate spatial distribution and quantitative data of the cavity, which solves the problems of the qualitative limitations and high missed detection rate of the traditional point drilling method.

[0020] (4) This invention ensures the accuracy of cavity location and volume results by setting up a secondary re-inspection and consistency comparison mechanism and initiating multi-source data fusion analysis. Furthermore, based on cavity parameters and geological support conditions, a risk model is used to achieve quantitative assessment of safety levels, providing direct and scientific decision-making basis for precise governance and realizing full-chain support from detection to risk control. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0023] Figure 2 This is a schematic diagram of the tunnel cross-section of the present invention.

[0024] Figure 3 This is a schematic diagram showing the connections of the various modules in the system of the present invention.

[0025] Reference numerals: 1. Top area of ​​the support; 2. Waist area of ​​the support; 3. Bottom area of ​​the support. Detailed Implementation

[0026] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figure 1 As shown, the present invention provides a method for detecting cavities in the initial support of tunnels in weak surrounding rock. The method includes: S1, performing position calibration based on tunnel contour point cloud data collected by a radar detection vehicle and tunnel contour standard data, and acquiring ground-penetrating radar detection data.

[0028] It should be noted that the contour point cloud data of the tunnel inner wall is collected by the lidar sensor on the radar detection vehicle. Specifically, the laser emitter emits a laser beam under the command of the control unit. Then, the scanning system drives the laser emitter to rotate or swing to change the emission direction of the laser beam, thereby scanning various positions on the tunnel inner wall. When the emitted laser beam hits the tunnel, it is reflected. The reflected laser signal is captured by the optical receiver, which converts the light signal into an electrical signal. The electrical signal is then amplified and filtered to obtain an effective electrical signal. By measuring the time difference between laser emission and reception, the target distance is calculated in combination with the speed of light. Combined with the laser emission angle recorded by the scanning system and the calculated target distance, the data is converted into three-dimensional coordinates using trigonometric functions. The three-dimensional coordinates of a large number of laser beams form dense contour point cloud data.

[0029] For example, the position calibration includes: extracting the tunnel contour of each detection point from the tunnel contour point cloud data, integrating the tunnel contours of each detection point to obtain the overall tunnel contour, and extracting the standard fluctuation range of the single-point tunnel contour and the standard range of the overall tunnel contour from the tunnel contour standard data.

[0030] The tunnel profile at each detection point is compared with the standard fluctuation range of the corresponding single-point tunnel profile, and the overall tunnel profile is compared with the standard range of the overall tunnel profile.

[0031] If the tunnel profile at each detection point is within its standard fluctuation range, and the overall tunnel profile completely encompasses the overall profile standard range, then the current position of the radar detection vehicle is maintained; otherwise, a positional deviation is determined.

[0032] If a positional deviation is determined, the actual center point sequence of the current tunnel is calculated, and this sequence is compared with the tunnel design axis. The offset of the radar detection vehicle in the lateral, longitudinal, and elevation directions is calculated, and the spatial coordinates of the radar detection vehicle are transformed and corrected based on the offset.

[0033] The specific process for calculating the actual center point sequence of the current tunnel is as follows: The point cloud data is divided into multiple equally spaced sections along the tunnel's longitudinal direction. For each section, all point cloud data within that section is extracted, and the contour curve of that section is fitted using methods such as ellipse fitting or least squares fitting. The center coordinates obtained from the fitting of each section are arranged in longitudinal order along the tunnel to obtain the actual center point sequence of the current tunnel.

[0034] It should be added that comparing the range of a single point contour can promptly detect deviations in local positions, preventing inaccurate positioning of individual detection points from affecting subsequent data. Comparing the overall contour coverage ensures the macroscopic accuracy of the tunnel contour, preventing large-scale detection errors caused by overall offset. This dual verification from local to global perspective allows for a more precise determination of whether the radar detection vehicle has any pose deviations. Subsequent coordinate transformation correction ensures the radar detection vehicle is in an accurate spatial position, laying the foundation for accurate analysis of subsequent ground-penetrating radar detection data. Ultimately, this improves the reliability of cavity detection results and ensures more accurate judgment of parameters such as the location and size of cavities in the initial support of tunnels in weak surrounding rock.

[0035] S2. Perform imaging analysis on the local electromagnetic wave characteristics in the ground-penetrating radar detection data to determine whether there are cavities. When a cavity is determined to exist, mark the corresponding location as a suspected cavity area and record the initial detection data.

[0036] For example, determining whether a void exists includes: obtaining the reflected wave amplitude, reflected wave phase, and reflected wave waveform from the local features of the electromagnetic wave.

[0037] Condition 1 is that the amplitude of the reflected wave is greater than a preset amplitude threshold for the reflected wave; condition 2 is that the phase of the reflected wave is reversed; and condition 3 is that the waveform of the reflected wave is hyperbolic.

[0038] The preset reflected wave amplitude threshold is a critical amplitude value set to distinguish strong reflection signals from cavities from background noise. It is obtained as follows: In a test section with similar geological and support conditions, a standard simulated cavity of known size and depth is artificially simulated. The area is scanned using radar equipment identical to that used in actual detection. Data is collected, and the maximum amplitude of reflected waves from each cavity and the background noise value of areas without cavities are extracted. After statistical analysis, the minimum amplitude value that covers the reflections of all known cavities is taken as the preset reflected wave amplitude threshold.

[0039] If condition 1 is true, and at least one of conditions 2 and 3 is true, then a void is determined to exist; otherwise, a void is determined not to exist.

[0040] It should be added that the reflected wave amplitude, reflected wave phase, and reflected wave waveform, among the local characteristics of electromagnetic waves, are chosen as the core criteria for cavity identification because: the reflected wave amplitude directly reflects the difference in dielectric constant between the two sides of the medium interface. The large difference in dielectric constant between the cavity and concrete or rock mass will induce strong reflection, with the amplitude significantly higher than that of the reflected wave from the surrounding homogeneous medium. Therefore, setting the reflected wave amplitude greater than a preset amplitude threshold as condition 1 can effectively screen energy anomaly areas and preliminarily indicate defects. Phase reversal of the reflected wave is a characteristic phenomenon of electromagnetic waves propagating from a medium with a high dielectric constant to a medium with a low dielectric constant. This change is usually not observed at the interface between concrete and rock mass. Therefore, setting phase reversal as condition 2 can specifically identify the cavity interface, improve the accuracy of identification, and distinguish between cavities and non-cavities. The hyperbolic shape of the reflected wave waveform is a typical spatial characteristic of cavities, formed by the change in the path difference between the radar antenna and the cavity reflector when the radar antenna moves. Therefore, setting the hyperbolic waveform as condition 3 can confirm the spatial attributes of the anomaly from its geometric shape, eliminating misjudgments from random noise or linear interference sources.

[0041] S3. For the suspected cavity area, the depth of the cavity is determined based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave velocity, and the longitudinal extension length and lateral width of the cavity are determined based on the radar image.

[0042] For example, determining the depth of the cavity includes: defining the boundary of a region with strong reflective signals in a radar image, and selecting measurement points at preset intervals along the longitudinal and transverse directions within that region.

[0043] The time difference between the surface reflected wave and the most significant cavity reflected wave in the waveform of each measuring point is extracted from the radar image, and the candidate cavity depth of each measuring point is calculated by combining the electromagnetic wave velocity.

[0044] The maximum value is selected from the candidate depths of the cavity at each measuring point.

[0045] Calculate the candidate depth gradient of the cavity at the core location where the maximum depth value is located, and analyze its decrease rate with increasing distance.

[0046] It should be added that the depth gradient is used to quantify the rate of change of depth around the core location of a cavity with distance. Its calculation requires combining the spatial distribution characteristics of the radar image and the morphological patterns of the cavity. Specifically, the method is as follows: taking the measuring point corresponding to the maximum depth value as the core location, first determine the lateral and longitudinal projection points of this core location in the radar image, and set these projection points as the origin of a plane coordinate system. This coordinate system is consistent with the actual spatial distribution of the tunnel and can directly correspond to the lateral and longitudinal extension directions of the cavity. Then, extract the plane coordinates of all measuring points around the core location and associate them with the candidate cavity depth values ​​of each measuring point. Based on the coordinate and depth data, first calculate the straight-line distance from each surrounding measuring point to the core location. Then, establish the relationship between distance and the candidate cavity depth through linear or polynomial fitting. The fitted relationship can intuitively reflect the trend of depth change with distance. Calculate the first derivative of this relationship to quantify how much depth changes with each unit increase in distance. The larger the absolute value of the derivative, the more drastic the change in depth with distance at that location, meaning this may be the boundary of a cavity; conversely, it indicates the interior of the cavity.

[0047] The decrease rate is used to determine the uniformity of the depth distribution in the core area of ​​a cavity, and thus to determine whether the depth value needs to be adjusted. The analysis method is as follows: with the core location as the center, the analysis interval is defined according to the preset gradient analysis radius. All depth gradient values ​​within the interval are extracted. First, the average gradient value is calculated. Then, the measurement point closest to the core and the measurement point farthest from the core within the interval are selected, and the distance difference between the two is calculated. According to the cavity morphology law that the depth gradient decreases with increasing distance, the absolute value of the ratio of the maximum change in gradient within the interval to the distance difference is used as the decrease rate. The larger the decrease rate, the faster the depth gradient decreases with increasing distance, which means that the depth distribution around the core location is uneven and the depth value needs to be corrected through robust statistical estimation. The smaller the decrease rate, the more uniform the depth distribution in the core area, and the maximum depth value can be directly used as the cavity depth.

[0048] If the decrease rate is less than the preset depth decrease rate threshold, the maximum depth value is taken as the depth of the cavity; otherwise, a robust statistical estimate is performed on the candidate depths of the cavities at each measuring point, and the estimated value is taken as the depth of the cavity.

[0049] It should be added that the preset depth reduction rate threshold is a preset critical value used to determine whether the depth distribution of the cavity is uniform. If the actual calculated reduction rate is less than the threshold, it indicates that the depth change around the core area of ​​the cavity is gradual and the depth distribution is relatively uniform. The maximum depth value can represent the overall depth of the cavity. If the actual reduction rate is greater than or equal to the threshold, it indicates that there is a significant abrupt change in the depth distribution of the cavity, and a more accurate cavity depth needs to be determined through robust statistical estimation.

[0050] The steps for obtaining the preset depth reduction rate threshold are as follows: A test section with geological conditions and support structure type consistent with the tunnel to be tested is selected. Standard cavities of different shapes and depths are artificially preset within the test section. A radar detection vehicle of the same model as the actual detection vehicle is used to collect radar detection data for each standard cavity using the same parameters. The depth gradient and reduction rate around the core location of each standard cavity are extracted. The reduction rate data of each standard cavity are statistically analyzed, and all samples whose maximum depth value accurately represents the actual depth of the standard cavity are selected. The maximum value of the reduction rate of these samples is taken as the initial preset depth reduction rate threshold.

[0051] The robust statistical estimation of the candidate cavity depth at each measuring point includes: screening outliers using the interquartile range method for the candidate cavity depth data at all measuring points, and calculating the first quartile of the depth data. With the third quartile The outlier determination interval is as follows: Depth data exceeding this range is identified as outliers and removed, retaining only the normal depth dataset. A threshold of 1.5 is used to screen for outliers, aiming to eliminate extreme values ​​that significantly deviate from the main data, thereby improving the reliability of depth estimation and ensuring that the calculation of hole depth is not affected by outliers.

[0052] Based on the dispersion of the normal depth dataset, an appropriate pruning ratio is set. A smaller ratio is used when the dispersion is small, and a larger ratio is used when the dispersion is large. After removing the extreme values ​​of each pruning ratio at both ends of the dataset, the arithmetic mean of the depth data in the remaining middle interval is calculated.

[0053] The calculated arithmetic mean is compared with the initial maximum depth value. If the deviation is less than a preset deviation threshold, the arithmetic mean is the robust statistical estimate. If the deviation is greater than the preset deviation threshold, the pruning ratio is adjusted, and the core statistic calculation and result verification steps are repeated until an estimate with acceptable deviation is obtained. The final estimate that meets the requirements is taken as the depth of the hole.

[0054] For example, determining the longitudinal extension length of the cavity includes: identifying the start and end boundaries of the strong reflective signal region along the longitudinal direction of the tunnel in the cross-section of the radar image, and using them as the initial longitudinal mileage range.

[0055] Based on a preset signal amplitude threshold, the signal amplitude within the initial longitudinal mileage range is filtered to determine the effective signal segment.

[0056] It should be added that determining the effective signal segment includes: within the initial longitudinal mileage range, collecting the amplitude data of the radar detection signal at a preset sampling interval, and associating the signal amplitude of each sampling point with its corresponding longitudinal mileage coordinate to form a dataset of mileage and amplitude.

[0057] The signal amplitude of each sampling point in the dataset is compared with the preset signal amplitude threshold. All sampling points with signal amplitudes greater than or equal to the preset signal amplitude threshold are selected, and the corresponding longitudinal mileage intervals of these sampling points are recorded.

[0058] The selected longitudinal mileage intervals are analyzed for continuity. If the longitudinal mileage interval between adjacent valid sampling points is less than the preset continuity threshold, they are merged into a continuous signal segment.

[0059] A minimum length threshold is set, and short continuous signal segments with a length less than the threshold are removed. The remaining continuous signal segments are the valid signal segments, and their corresponding longitudinal mileage range is used as the key analysis range for subsequent cavity detection.

[0060] The preset signal amplitude threshold is a critical amplitude value used to distinguish effective hole reflections from background noise in radar signals. When the amplitude of the radar detection signal is greater than or equal to this threshold, it indicates that the signal has the energy characteristics of hole reflection; otherwise, it is determined to be an invalid signal with insufficient energy and is not included in subsequent hole analysis.

[0061] The steps for obtaining the preset signal amplitude threshold are as follows: Select a test section with geological conditions and support structure parameters consistent with the tunnel to be tested, and artificially preset standard cavities of different sizes and filling states within the test section. Use radar equipment of the same model as the actual test equipment to collect the reflection signals of each standard cavity with the same parameters, and record the amplitude data of the reflection signal of each standard cavity. Perform statistical analysis on the collected standard cavity reflection signal amplitudes, remove outliers caused by equipment noise and operational errors, and take the minimum value of all valid amplitudes as the preset signal amplitude threshold.

[0062] Within the effective signal segment, the continuity of the phase axis of the cavity reflected wave is traced longitudinally. Combined with the cavity candidate depth at each measuring point, the cavity candidate depth at each measuring point within the effective signal segment is verified to be greater than the non-cavity depth threshold. The non-cavity depth threshold is the critical depth value that distinguishes between effective cavities and normal support thickness, and its value is 85% of the initial support design thickness. According to the relevant standards regarding the allowable deviation of support thickness, it is usually ±15%. Therefore, taking 85% can effectively cover the allowable negative deviation and reserve a certain margin.

[0063] The mileage range where the phase axis is continuous and the candidate depth of the cavity is consistently greater than the non-cavity depth threshold is taken as the true longitudinal extension range of the cavity.

[0064] The longitudinal extension length of the cavity is obtained by calculating the mileage difference of the actual longitudinal extension range.

[0065] For example, determining the lateral width of the cavity includes: obtaining a radar image time slice corresponding to the maximum depth value from the radar image, and then extracting the signal amplitude curve distributed laterally along the tunnel.

[0066] Identify the first continuous region in the signal amplitude curve where the amplitude exceeds a preset threshold.

[0067] It should be added that the preset threshold refers to a critical value of signal amplitude set in advance to distinguish cavity reflection signals from non-cavity interference signals, based on the characteristics of ground-penetrating radar detection signals and the requirements for calculating the lateral width of cavities. After extracting the tunnel lateral signal amplitude curve, this threshold is used to filter out effective signal areas with cavity reflection characteristics, avoiding misjudgments of the cavity lateral width caused by interference signals, and ensuring that the first continuous area calculated subsequently accurately corresponds to the actual lateral distribution range of the cavity.

[0068] The specific method for obtaining the preset threshold is as follows: From the ground-penetrating radar detection data of the tunnel to be detected, select the normal support area that has been preliminarily determined to be free of cavity interference. Obtain the lateral signal amplitude data of the radar image time slice at the same depth level as the maximum depth value of the cavity in the normal support area. Extract the maximum value of this set of data as the preset threshold, thereby ensuring that the maximum interference signal amplitude of the normal support area does not exceed the threshold, and thus accurately filtering out the reflected signals of the cavity area that exceed the normal interference range, avoiding deviations in the calculation of the lateral width.

[0069] S4. Based on the depth, longitudinal extension length, and lateral width of the cavity, calculate the initial volume using a pre-set cavity shape graphic library. The cavity shape graphic library includes standard geometric shapes such as spheres, ellipsoids, cylinders, and prisms, along with their volume calculation formulas.

[0070] For example, the calculation of the initial detection volume includes: constructing an initial three-dimensional envelope representing the core area of ​​the cavity based on the depth, longitudinal extension length and lateral width of the cavity.

[0071] Identify the signal transition zone surrounding the core area of ​​the cavity in the radar image, wherein the signal transition zone satisfies the condition that the reflected wave amplitude is between that of the core area of ​​the cavity and the background noise of the surrounding rock, and the phase axis is discontinuous.

[0072] Based on the signal transition zone range, and in conjunction with the support structure type and surrounding rock grade, the potential stripping range expansion coefficient in each dimension is determined.

[0073] It should be added that the potential stripping range expansion coefficient refers to the coefficient set in the three dimensions of depth, longitudinal direction, and transverse direction based on the fuzzy characteristics of the stripping boundary reflected by the radar detection signal transition zone, combined with the stripping resistance of the support structure and the difference in the stability of the surrounding rock. It is used to expand the initial stripping range corresponding to the signal transition zone into a potential stripping range that is more in line with the actual project. Its core function is to avoid the omission of stripping range due to the fuzzy signal boundary. In particular, it is used to address the problem of the stripping boundary and normal support signal superimposed in the initial support of weak surrounding rock tunnels. By expanding the coefficient, the actual stripping area can be accurately covered.

[0074] The determination of the potential stripping range expansion coefficient in each dimension includes: setting the basic expansion coefficient according to the structural type of the initial support of the tunnel to be tested, combined with the difference in the anti-stripping ability of different support structures. For example, if it is a single support of shotcrete, the depth foundation coefficient and the transverse foundation coefficient are set upward according to the signal transition zone ratio, and the longitudinal foundation coefficient is set according to the signal transition zone ratio. The value of the foundation coefficient must ensure that it covers the fuzzy range corresponding to the signal transition zone, and the foundation coefficient of each dimension is greater than 1.

[0075] The initial three-dimensional envelope is expanded outward based on the expansion coefficient to form a comprehensive three-dimensional envelope, wherein the comprehensive three-dimensional envelope includes a core region and a potential stripping region.

[0076] The comprehensive three-dimensional envelope is matched with a preset cavity shape graphic library for similarity. If a standard geometric object with a similarity of up to a preset threshold is matched, the volume formula corresponding to the standard geometric object is used to calculate the initial detection volume.

[0077] The similarity matching is achieved by calculating the Euclidean distance between the shape feature vectors of the integrated 3D envelope and the standard geometric object. If the Euclidean distance is less than a preset threshold, the match is considered successful.

[0078] It should be added that the standard geometry of the preset threshold is a series of pre-constructed basic sets with specific geometric shapes and size ranges, used to provide quantitative reference standards for the actual detected target objects such as holes. Its construction is based on common regular geometric shapes, such as cuboids, spheres, cylinders, and cones. These standard geometries each correspond to different application scenarios; for example, a cuboid can simulate the shape of a hole behind a wall, a sphere can correspond to a circular hole, and a cylinder can represent a tubular hole or a hole inside a columnar structure.

[0079] The dimensional parameter ranges are determined based on actual engineering needs, relevant testing standards, and statistical analysis of extensive historical data. For example, in the scenario of detecting voids in the initial support of a tunnel, the length, width, and height threshold ranges for a cuboid are set by referring to the structural dimensions in the tunnel design drawings, the allowable construction error range, and the common void size distribution found in previous inspections. For a sphere, a reasonable threshold range for its radius is determined. Simultaneously, considering the differences in characteristics of different structural parts, such as the arch crown, arch waist, and sidewalls, the shape and size tendencies of voids differ due to different stress conditions and construction processes, and the preset thresholds of the standard geometry are adjusted accordingly. Using these preset threshold standard geometries, during actual inspection, the actual geometric parameters of the detected voids can be compared with these parameters to quickly determine whether the void's geometric characteristics exceed the normal range, thereby assisting in assessing the structural risk level.

[0080] If a standard geometry with a similarity score below the preset threshold is matched, the composite 3D envelope is divided into sub-geometry and re-matched with the graphics library. If a match is successful, the volume of each sub-geometry is calculated and summed. If a match fails, the volume of each sub-geometry is calculated using numerical integration, and then the volumes of each sub-geometry are summed to obtain the initial detection volume.

[0081] S5. Conduct a second inspection of the suspected cavity area, compare and analyze the consistency between the second inspection data and the initial inspection data, determine the actual location coordinates and actual volume of the cavity, assess the structural risk level accordingly, and generate a cavity detection report.

[0082] For example, determining the actual location coordinates and actual volume of the cavity includes: spatially registering the re-inspection data with the initial inspection data, wherein spatial registration is performed to ensure that the two inspection data are in the same coordinate system.

[0083] Extract the re-inspection depth, re-inspection longitudinal extension length, and re-inspection transverse width of the cavity from the re-inspection data, and calculate the re-inspection volume.

[0084] The initial inspection volume and the re-inspection volume are compared for consistency, and the relative volume error between the two is calculated.

[0085] If the relative error of the volume is less than or equal to the preset error threshold, the average value of the position coordinates and volume in the initial inspection and re-inspection results will be used as the actual position coordinates and actual volume, respectively. Otherwise, the third re-inspection and manual judgment process will be initiated to determine the actual position coordinates and actual volume of the cavity.

[0086] It should be added that the supplementary verification process specifically includes: adjusting the detection parameters of the radar detection vehicle, conducting a third targeted re-inspection of the suspected cavity area, collecting re-inspection radar image data and corresponding tunnel contour point cloud data, repeating the calculation logic of steps 3 and 4, and obtaining the cavity location coordinates and volume data of the third re-inspection.

[0087] The location coordinates and volume data of the three sets of cavities from the initial inspection, re-inspection, and third re-inspection were compared, the dispersion of the three sets of data was calculated, and the three sets of radar images were superimposed to compare the morphology of the strong reflection signal of the cavity and the consistency of the continuity of the phase axis in the three inspections, and the difference areas were marked.

[0088] The judgment is made by professionals who combine the radar waveform characteristics of the different areas, tunnel construction records, and the equipment status and environmental parameters of the three tests.

[0089] If the dispersion of the third re-inspection data and a certain previous detection data is less than the preset correlation threshold, and there is no obvious interference after manual judgment, then the mean of the two sets of consistent data is used as the actual location coordinates and actual volume of the cavity. If there are differences in all three sets of data, but the core area of ​​the cavity can be locked through waveform features, then the actual location coordinates and actual volume are recalculated based on the cavity boundary determined by manual judgment.

[0090] Please see Figure 2 As shown, for example, the assessment of structural risk level includes: determining whether the cavity is located in the top, middle or bottom region of the support based on the actual location coordinates of the cavity, and querying the preset initial location reference coefficients for each region to obtain the initial location reference coefficients of the cavity.

[0091] It should be added that the preset initial position reference coefficients for each region refer to coefficients pre-set to calibrate the initial positions of the corresponding detection parts of the radar detection vehicle, based on the design position characteristics of specific structural parts in different regions of the tunnel support (top, waist, and bottom). These coefficients transform the theoretical positions of specific parts such as the top, waist, and bottom of the support in the tunnel design drawings into actual detection coordinate references that the radar detection system can recognize. This ensures that the radar's starting point for detecting the position of different regions and corresponding specific parts of the support is consistent with the design position, avoiding positional deviations caused by differences in the structure of regions and parts.

[0092] The method for obtaining the preset initial position reference coefficients for each region is as follows: referring to the tunnel construction design drawings, the region is first divided into the top, waist and bottom of the support, and the key structural parts to be detected in each region are determined accordingly. Then, the design coordinate parameters such as the lateral offset, elevation difference and longitudinal mileage reference of the corresponding specific parts in each region are extracted from the drawings. Finally, according to the coordinate adaptation rules of the radar detection system, the extracted coordinate parameters are converted into the initial position reference coefficients of each region that the system can recognize.

[0093] The vertical height of the cavity within its area and the vertical height from the top to the bottom of the area are extracted from the spatial coordinates corresponding to the ground-penetrating radar data, and the ratio of the two is used as the position correction coefficient.

[0094] The position coefficient is obtained by multiplying the initial position reference coefficient of the cavity with the position correction coefficient.

[0095] The actual volume of the cavity is normalized to obtain a normalized volume coefficient. Then, the normalized volume coefficient and the location coefficient are weighted and fused to calculate the structural safety risk value.

[0096] It should be added that the normalization process includes: taking the minimum value of the design parameters of the initial support of the tunnel to be tested and the void volume threshold that needs to be focused on in the engineering specifications as the benchmark value, and taking the ratio of the actual void volume to the benchmark value as the normalized volume coefficient.

[0097] The structural safety risk value is compared with the risk threshold range corresponding to each structural risk level to obtain the structural risk level of the void.

[0098] It should be added that the weighted fusion calculation of structural safety risk values ​​aims to comprehensively consider the contributions of normalized volume coefficients and location coefficients to structural safety risk. On one hand, the weight allocation reflects the relative importance of different factors. Weights can be set based on engineering safety standards, practical engineering experience, or obtained through experimental data. For example, historical data on the impact of normalized volume coefficients, location coefficients, and their corresponding structural safety risk are first collected, and their correlation coefficients with structural safety risk are calculated. Regression analysis or logistic regression analysis is then used to determine the contribution of each factor to structural safety risk. After normalization, the contribution is converted into weights, and the sum of the weights is 1.

[0099] On the other hand, weighted fusion achieves the quantitative integration of multi-dimensional information. Its calculation formula can be expressed as: ,in This represents the structural safety risk value. The normalized volume coefficient, For position coefficients, and These are the weights for the normalized volume coefficient and the location coefficient, respectively. In this embodiment, as an optional weighting scheme, the normalized volume coefficient weight can be set to 0.7, and the location coefficient weight to 0.3. This setting is based on engineering experience, emphasizing the dominant influence of cavity size on safety risks while also taking into account the adverse effects of its location.

[0100] Please see Figure 3 As shown, the present invention provides a cavity detection system for the initial support of tunnels in weak surrounding rock. The system includes: a point cloud calibration module, a cavity identification module, a parameter extraction module, a volume calculation module, and a re-inspection report module.

[0101] In the above, the point cloud calibration module is connected to the hole identification module and the parameter extraction module, the parameter extraction module is connected to the hole identification module and the volume calculation module, and the volume calculation module is also connected to the re-inspection report module.

[0102] The point cloud calibration module performs position calibration based on the tunnel contour point cloud data collected by the radar detection vehicle and the tunnel contour standard data, and acquires ground-penetrating radar detection data.

[0103] The cavity identification module performs imaging analysis on the local electromagnetic wave features in the ground-penetrating radar detection data to determine whether a cavity exists. When a cavity is determined to exist, the corresponding location is marked as a suspected cavity area, and the initial detection data is recorded.

[0104] The parameter extraction module determines the depth of the suspected cavity based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave velocity, and determines the longitudinal extension length and lateral width of the cavity based on the radar image.

[0105] The volume calculation module calculates the initial volume based on the depth, longitudinal extension length, and lateral width of the cavity, combined with a preset cavity shape graphic library.

[0106] The re-inspection report module performs a second re-inspection on the suspected cavity area, compares and analyzes the consistency between the re-inspection data and the initial inspection data, determines the actual location coordinates and actual volume of the cavity, assesses the structural risk level accordingly, and generates a cavity detection report.

[0107] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0108] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0109] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0110] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0111] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting voids in the initial support of tunnels in weak surrounding rock, characterized in that: The method includes: S1. Based on the tunnel contour point cloud data collected by the radar detection vehicle and the tunnel contour standard data, position calibration is performed, and ground-penetrating radar detection data is obtained. The position calibration includes: The tunnel contour of each detection point is extracted from the tunnel contour point cloud data, the tunnel contour of each detection point is integrated to obtain the overall tunnel contour, and the standard fluctuation range of the single-point tunnel contour and the standard range of the overall tunnel contour are extracted from the tunnel contour standard data. The tunnel profile at each detection point is compared with the standard fluctuation range of the corresponding single-point tunnel profile, and the overall tunnel profile is compared with the standard range of the overall tunnel profile. If the tunnel profile at each detection point is within its standard fluctuation range and the overall tunnel profile completely encompasses the overall profile standard range, then the current position of the radar detection vehicle is maintained; otherwise, a positional deviation is determined. If a positional deviation is determined, the actual center point sequence of the current tunnel is calculated and compared with the tunnel design axis. The offset of the radar detection vehicle in the lateral, longitudinal and elevation directions is calculated, and the spatial coordinates of the radar detection vehicle are transformed and corrected based on the offset. S2. Perform imaging analysis on the local electromagnetic wave characteristics in the ground-penetrating radar detection data to determine whether there are cavities. When a cavity is determined to exist, mark the corresponding location as a suspected cavity area and record the initial detection data. S3. For the suspected cavity area, the depth of the cavity is determined based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave speed, and the longitudinal extension length and lateral width of the cavity are determined based on the radar image. S4. Calculate the initial inspection volume based on the depth, longitudinal extension length, and lateral width of the cavity, combined with a preset cavity shape graphic library. S5. Conduct a second inspection of the suspected cavity area, compare and analyze the consistency between the second inspection data and the initial inspection data, determine the actual location coordinates and actual volume of the cavity, assess the structural risk level accordingly, and generate a cavity detection report.

2. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: The determination of whether a void exists includes: The amplitude, phase, and waveform of the reflected wave are obtained from the local features of the electromagnetic wave. Condition 1 is that the amplitude of the reflected wave is greater than a preset amplitude threshold for the reflected wave; condition 2 is that the phase of the reflected wave is reversed; and condition 3 is that the waveform of the reflected wave is hyperbolic. If condition 1 is true, and at least one of conditions 2 and 3 is true, then a void is determined to exist; otherwise, a void is determined not to exist.

3. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: Determining the depth of the cavity includes: Define the boundary of the strong reflection signal area in the radar image, and select each measuring point in the longitudinal and transverse directions within the area at a preset interval; Extract the time difference between the surface reflected wave and the most significant cavity reflected wave from the waveform of each measuring point in the radar image, and calculate the candidate cavity depth of each measuring point by combining the electromagnetic wave velocity. The maximum value is selected from the candidate depths of the cavity at each measuring point and taken as the maximum depth value. Calculate the candidate depth gradient of the cavity at the core location where the maximum depth value is located, and analyze its decrease rate with increasing distance. If the decrease rate is less than the preset depth decrease rate threshold, the maximum depth value is taken as the depth of the cavity; otherwise, a robust statistical estimate is performed on the candidate depths of the cavities at each measuring point, and the estimated value is taken as the depth of the cavity.

4. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: The determination of the longitudinal extension length of the cavity includes: In the profile of the radar image, the start and end boundaries of the region with strong reflection signals are identified along the longitudinal direction of the tunnel and used as the initial longitudinal mileage range. Based on a preset signal amplitude threshold, the signal amplitude within the initial longitudinal mileage range is filtered to determine the effective signal segments; Within the effective signal segment, the continuity of the phase axis of the cavity reflected wave is traced longitudinally, and the cavity candidate depth at each measuring point is combined to verify whether the cavity candidate depth at each measuring point within the effective signal segment is greater than the non-cavity depth threshold. The mileage range where the phase axis is continuous and the candidate depth of the cavity is consistently greater than the non-cavity depth threshold is taken as the true longitudinal extension range of the cavity. The longitudinal extension length of the cavity is obtained by calculating the mileage difference of the actual longitudinal extension range.

5. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: The determination of the lateral width of the cavity includes: The radar image time slice corresponding to the maximum depth value is obtained from the radar image, and then the signal amplitude curve distributed laterally along the tunnel is extracted. Identify the first continuous region in the signal amplitude curve where the amplitude exceeds a preset threshold; Calculate the projected distance of the first continuous region in the transverse direction of the tunnel, and use it as the transverse width of the cavity.

6. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: The calculation of the initial detection volume includes: Based on the depth, longitudinal extension length, and lateral width of the cavity, an initial three-dimensional envelope representing the core region of the cavity is constructed. Identify the signal transition zone surrounding the cavity core region in the radar image; Based on the signal transition zone range, and in combination with the support structure type and surrounding rock grade, determine the potential stripping range expansion coefficient in each dimension. The initial three-dimensional envelope is expanded outward based on the expansion coefficient to form a comprehensive three-dimensional envelope; The comprehensive three-dimensional envelope is matched with a preset cavity shape graphic library for similarity. If a standard geometric body with a similarity of up to a preset threshold is matched, the volume formula corresponding to the standard geometric body is used to calculate the initial detection volume. If the matching fails, the composite 3D envelope is divided into sub-geometry and rematched with the graphics library. If the matching is successful, the volume of each sub-geometry is calculated and then summed. If the matching fails, the volume of each sub-geometry is calculated using numerical integration, and then the volumes of each sub-geometry are summed to obtain the initial detection volume.

7. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: Determining the actual location coordinates and actual volume of the cavity includes: Spatial registration is performed between the re-inspection data and the initial inspection data; Extract the re-inspection depth, re-inspection longitudinal extension length, and re-inspection transverse width of the cavity from the re-inspection data, and calculate the re-inspection volume; The initial inspection volume and the re-inspection volume are compared for consistency, and the relative volume error between the two is calculated. If the relative error of the volume is less than or equal to the preset error threshold, the average value of the position coordinates and volume in the initial inspection and re-inspection results will be used as the actual position coordinates and actual volume, respectively. Otherwise, the third re-inspection and manual judgment process will be initiated to determine the actual position coordinates and actual volume of the cavity.

8. The method for detecting voids in the initial support of a tunnel in weak surrounding rock according to claim 1, characterized in that: The risk levels of the assessed structure include: Based on the actual location coordinates of the cavity, determine whether it is located in the top, middle or bottom area of ​​the support, and query the preset initial location reference coefficients for each area to obtain the initial location reference coefficients of the cavity. The vertical height of the cavity within its area and the vertical height from the top to the bottom of the area are extracted from the spatial coordinates corresponding to the ground-penetrating radar detection data, and the ratio of the two is used as the position correction coefficient. The initial position reference coefficient and the position correction coefficient of the cavity are multiplied to obtain the position coefficient; The actual volume of the cavity is normalized to obtain a normalized volume coefficient. Then, the normalized volume coefficient and the location coefficient are weighted and fused to calculate the structural safety risk value. The structural safety risk value is compared with the risk threshold range corresponding to each structural risk level to obtain the structural risk level of the void.

9. A system for detecting voids in the initial support of tunnels in weak surrounding rock, characterized in that: The system includes: The point cloud calibration module performs position calibration based on the tunnel outline point cloud data collected by the radar detection vehicle and the tunnel outline standard data, and acquires ground-penetrating radar detection data. The position calibration includes: The tunnel contour of each detection point is extracted from the tunnel contour point cloud data, the tunnel contour of each detection point is integrated to obtain the overall tunnel contour, and the standard fluctuation range of the single-point tunnel contour and the standard range of the overall tunnel contour are extracted from the tunnel contour standard data. The tunnel profile at each detection point is compared with the standard fluctuation range of the corresponding single-point tunnel profile, and the overall tunnel profile is compared with the standard range of the overall tunnel profile. If the tunnel profile at each detection point is within its standard fluctuation range and the overall tunnel profile completely encompasses the overall profile standard range, then the current position of the radar detection vehicle is maintained; otherwise, a positional deviation is determined. If a positional deviation is determined, the actual center point sequence of the current tunnel is calculated and compared with the tunnel design axis. The offset of the radar detection vehicle in the lateral, longitudinal and elevation directions is calculated, and the spatial coordinates of the radar detection vehicle are transformed and corrected based on the offset. The cavity identification module performs imaging analysis on the local electromagnetic wave features in the ground-penetrating radar detection data to determine whether cavities exist. When a cavity is determined to exist, the corresponding location is marked as a suspected cavity area, and the initial detection data is recorded. The parameter extraction module determines the depth of the cavity based on the time difference between the surface reflected wave and the cavity reflected wave and the electromagnetic wave velocity for the suspected cavity area, and determines the longitudinal extension length and lateral width of the cavity based on the radar image. The volume calculation module calculates the initial volume based on the depth, vertical extension length, and horizontal width of the cavity, combined with a preset cavity shape graphic library. The re-inspection report module performs a second re-inspection on the suspected cavity area, compares and analyzes the consistency between the re-inspection data and the initial inspection data, determines the actual location coordinates and actual volume of the cavity, assesses the structural risk level accordingly, and generates a cavity detection report.

Images