Tunnel cave water inflow prediction analysis method and system based on ground penetrating radar

Abstract

The present application belongs to the technical field of tunnel cave data analysis, and in particular relates to a tunnel cave water inflow prediction analysis method and system based on a geological radar. The method first extracts amplitude intensity, hyperbolic shape and disorder degree from 3D radar volume data, then cooperatively identifies abnormal reflection regions, then for the abnormal cave sections, extracts abnormal parameters to comprehensively calculate a water inflow risk index, simultaneously calculates an effective water storage volume according to the abnormal body parameters, and combines radar data to identify the current water content, then combines the water inflow risk index, water content and water storage volume to calculate the estimated water inflow of each section, and finally comprehensively predicts and outputs the overall estimated water inflow of the tunnel cave according to a preset rule. The present application effectively improves the accuracy of the water inflow of the tunnel cave, can provide reliable data support for water inflow prevention and control decisions during tunnel construction, effectively reduces the risk of water inflow accidents, and ensures construction safety and efficiency.

Description

Technical Field

[0001] This invention belongs to the field of tunnel and karst cave data analysis technology, specifically, it relates to a method and system for predicting and analyzing the water inflow of tunnels and karst caves based on ground-penetrating radar. Background Technology

[0002] Groundwater in tunnels can dissolve rock strata over a long period of time, creating karst caves. Water inrush from these caves is the most typical and extremely dangerous hazard, making the prediction of their inrush volume crucial.

[0003] Existing methods for predicting and analyzing water inflow rely heavily on 2D radar detection or single parameters, which are insufficient to cover the three-dimensional spatial characteristics of karst caves. They are prone to overlooking small-scale anomalies or misjudging anomalous areas, leading to distortion of the basic data for subsequent water inflow calculations. The following problems also exist: 1. Water inflow is estimated by using rock mass properties or a single anomalous area alone, without considering the specific distribution characteristics of the anomalies and the risk of water inflow, resulting in a large deviation in the prediction results.

[0004] 2. Currently, the karst cave is treated as a whole in the calculation of water inflow, without taking into account the spatial location of the karst cave section and its dynamic impact on the water inflow. This makes the water inflow calculation unable to adapt to the actual differences in different heights and distances from different excavation faces, resulting in poor reference value of the calculation results. Summary of the Invention

[0005] In view of this, in order to solve the above problems, a method and system for predicting and analyzing water inflow in tunnels and karst caves based on ground-penetrating radar is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: This invention provides a method for predicting and analyzing the water inflow of tunnel karst caves based on ground-penetrating radar. The method includes: dividing the tunnel karst cave into several karst cave segments from top to bottom according to a preset height, and performing 3D radar detection on each karst cave segment to obtain 3D radar volume data, and marking the height of the karst cave segment.

[0007] Based on the 3D radar volume data, amplitude intensity, hyperbolic shape, and degree of disorder are obtained to identify areas of abnormal reflection.

[0008] For karst cave sections with reflective anomalies, the area of ​​the anomaly region, the number of anomalies, the spatial location and cross-sectional area of ​​each anomaly, and its estimated length in the tunnel excavation direction are extracted, and the water inrush risk index of each karst cave section is calculated in a comprehensive manner.

[0009] The effective water storage volume of each anomaly is calculated using a water-rich anomaly volume estimation model based on the cross-sectional area of ​​each anomaly and the estimated length in the tunnel excavation direction.

[0010] The estimated water inflow of each karst cave section is calculated based on the water inflow risk index and effective water storage volume of each section. The estimated water inflow of the tunnel karst cave is then comprehensively predicted according to the preset prediction rules and output.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention identifies the abnormal reflection area by combining the three-dimensional parameters of amplitude intensity, hyperbolic shape and disorder degree, which provides accurate positioning and range basis for the identification of effective water storage volume meter, avoids attribute quantification deviation caused by boundary ambiguity, and thus ensures that the subsequent identification results are related to the actual geological conditions.

[0012] (2) The present invention corrects and calculates the estimated water inflow by the spatial location of the karst cave section, and fully considers the gravity release law of water, so that the static water accumulation prediction is more in line with reality. At the same time, it accurately distinguishes the dynamic water inflow risk source and the static water accumulation area, which makes it easier to lock the core target of prevention and control, and further greatly reduces the prediction deviation of water inflow.

[0013] (3) This invention calculates the distance risk degree by combining the vertical distance between the anomaly and the excavation face, so that the water inrush risk index is adapted to the real-time risk changes during the tunnel excavation process. It fully considers the influence of the water inrush risk index, height difference and anomaly distribution on the water inrush volume in the karst section, and ensures the prediction accuracy and predictability in complex karst group scenarios.

[0014] (4) This invention estimates the inflow by correcting the effective water storage volume based on the proportion of high-risk karst cave sections, thereby achieving dynamic adaptation between the risk proportion and the effective water storage volume. This ensures that the effective water storage volume is always associated with the inflow risk, and thus achieves dynamic matching between the effective water storage volume and the actual inflow probability. This results in more reliable predictions, which facilitates targeted and timely risk management in the future. Attached Figure Description

[0015] 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.

[0016] Figure 1 This is a schematic diagram of the overall implementation process of the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the calculation process of the inrush risk index of this invention.

[0018] Figure 3 This is a schematic diagram of the prediction process of the prediction rule of the present invention.

[0019] Figure 4This is a schematic diagram illustrating the process for selecting the baseline inflow rate according to the present invention.

[0020] Figure 5 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0021] 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.

[0022] Please see Figure 1 As shown, a method for predicting and analyzing water inflow in tunnel karst caves based on ground-penetrating radar is described. This method includes: dividing the tunnel karst cave into several cave segments from top to bottom according to a preset height, and conducting 3D radar detection on each cave segment to obtain 3D radar volume data, and marking the height of the cave segment.

[0023] It should be noted that the above-mentioned preset height can be obtained through industry experience. For example, in order to reduce the pressure of data analysis, the tunnel can be divided into three equal parts according to its height direction, that is, one-third of the tunnel excavation height can be used as the preset height.

[0024] Based on the 3D radar volume data, amplitude intensity, hyperbolic shape, and degree of disorder are obtained to identify areas of abnormal reflection.

[0025] Specifically, the process of obtaining the amplitude intensity includes: selecting the reflected wave signal waveform of each sampling point within the corresponding depth range of each cave segment, extracting the amplitude from the waveform, and obtaining the relative amplitude value after performing minimum-maximum normalization processing on the amplitude.

[0026] Calculate the mean of the relative amplitude values ​​corresponding to each sampling point, and then filter out the maximum relative amplitude value.

[0027] The ratio of the number of relative amplitude values ​​exceeding the preset amplitude threshold in each karst cave section to the total number of relative amplitude values ​​in the corresponding karst cave section is used as the amplitude ratio.

[0028] The amplitude intensity is output by comparing the amplitude ratio with each preset amplitude ratio threshold.

[0029] It should be noted that the aforementioned preset amplitude ratio thresholds include a first preset amplitude ratio threshold and a second preset amplitude ratio threshold, and the first preset amplitude ratio threshold is greater than the second preset amplitude ratio threshold.

[0030] It should be added that the specific steps of comparing the amplitude ratio with each preset amplitude ratio threshold and then outputting the corresponding amplitude intensity include: if the amplitude ratio is greater than the first preset amplitude ratio threshold, the average amplitude is taken as the amplitude intensity.

[0031] If the amplitude ratio is less than the second preset amplitude ratio threshold, the maximum amplitude value will be used as the amplitude intensity.

[0032] If the amplitude ratio is less than the first preset amplitude ratio threshold but greater than the second preset amplitude ratio threshold, calculate the difference between the maximum amplitude and the average amplitude. When the difference is less than the preset difference threshold, use the average amplitude as the amplitude intensity.

[0033] When the difference is greater than or equal to the preset difference threshold, the average of the maximum amplitude and the average amplitude is taken as the amplitude intensity.

[0034] It should be noted that the first preset amplitude ratio threshold and the second preset amplitude ratio threshold mentioned above correspond to the scenarios where there are a lot of strong reflection signals and the average value is used to represent the amplitude intensity, and the scenarios where there are very few strong reflection signals and the maximum value is used to represent the amplitude intensity, respectively.

[0035] In practice, the first preset amplitude ratio threshold and the second preset amplitude ratio threshold can be obtained by performing 3D radar detection on multiple normal segments, calculating the amplitude ratio of each normal segment, calculating the mean and standard deviation of the amplitude ratio of the normal segments, adding the mean to twice the standard deviation to obtain the first preset amplitude ratio threshold, and finally subtracting the mean from the standard deviation to obtain the second preset amplitude ratio threshold.

[0036] Furthermore, the aforementioned preset difference threshold is obtained based on industry experience; for example, 0.1 is used as the preset difference threshold.

[0037] It should be added that by combining the first preset amplitude ratio threshold and the second preset amplitude ratio threshold, the amplitude ratio range is divided. The mean value is taken when the ratio is high and the maximum value is taken when the ratio is low, which adapts to the differences in water-rich distribution and makes the amplitude intensity more in line with the actual geological characteristics. At the same time, the difference between the maximum value and the mean value is used to judge, so that key risk information is not lost, the overall characteristics of the region are reflected, and more accurate quantitative basis is provided for subsequent identification of abnormal areas.

[0038] Specifically, the process of obtaining the degree of disorder includes: taking waveform profiles of the radar detection area in the tunneling direction, lateral direction, and height direction respectively, and extracting the amplitude value and phase of each sampling point in the radar detection area.

[0039] Calculate the amplitude difference between all adjacent sampling points, and then calculate the mean amplitude difference between adjacent sampling points. Simultaneously extract the maximum amplitude value .

[0040] Similarly, the average phase difference between adjacent sampling points can be calculated using the same method as the average amplitude difference. .

[0041] The waveform continuity index of the radar detection area is calculated based on the mean amplitude difference, the maximum amplitude value, and the mean phase difference. The complement of the exponent is taken as the degree of disorder.

[0042] It should be noted that the formula for calculating the waveform continuity index of the radar detection area is as follows: .

[0043] In the formula, This is the reference value for the maximum phase difference. This is the amplitude continuity weight. This is the phase continuity weight.

[0044] When the tunnel cavity is composed of limestone, due to the high density of limestone, the amplitude continuity responds more directly to abrupt changes in the medium. Furthermore, the high phase stability of limestone makes the phase continuity response to anomalies weaker than that of amplitude. Therefore, in limestone tunnel cavity detection, amplitude continuity contributes more to the identification of water-rich anomalies and is thus given a higher weight. Phase continuity, as an auxiliary reference, has a relatively lower weight, ensuring that the index calculation prioritizes the core anomaly signal. Specifically… The example value is 0.55. The example value is 0.45.

[0045] The amplitude continuity deviation coefficient represents the degree of dispersion of amplitude changes between adjacent sampling points within the radar detection area. The larger the amplitude continuity deviation coefficient, the more significant the amplitude difference between adjacent sampling points, that is, the larger the mean amplitude difference between adjacent sampling points, the weaker the continuity of the waveform in the amplitude dimension, which reflects the poorer amplitude continuity.

[0046] The phase continuity deviation coefficient characterizes the degree of disorder in the phase change of adjacent sampling points within the radar detection area. The larger the phase continuity deviation coefficient, the more drastic the phase jump between adjacent sampling points, that is, the larger the average phase difference between adjacent sampling points, the worse the consistency of the waveform in the phase dimension, and the worse the phase continuity.

[0047] It should be noted that the process of identifying the above-mentioned abnormal reflection areas includes: calculating the average amplitude intensity of the normal reflection of the surrounding rock in each karst cave section, and screening out areas with amplitude intensity greater than the preset amplitude intensity threshold, which are recorded as strong amplitude areas.

[0048] It should be noted that the above-mentioned preset amplitude intensity threshold can be determined based on industry experience. For example, the preset amplitude intensity threshold is 1.5 times the average amplitude intensity.

[0049] Areas with symmetrical hyperbolic reflections of the same shape in three or more consecutive detection slices along the tunnel excavation direction within the strong amplitude region are marked as potential anomaly areas.

[0050] Extract the degree of disorder in each potential abnormal region. If the degree of disorder is greater than the threshold for judging abnormality, the region is judged as a reflection abnormal region.

[0051] The setting of the abnormal threshold can be achieved by first extracting the normal area of ​​the surrounding rock without strong amplitude and hyperbola in the karst cave section, and calculating its disorder level as the benchmark data. Then, the mean and standard deviation of the benchmark data are calculated. Finally, the mean and the standard deviation at a certain multiple are summed to output the abnormal threshold. The standard deviation multiple can be determined based on industry experience, and for example, a value of 2 can be taken.

[0052] For karst cave sections with reflective anomalies, the area of ​​the anomaly region, the number of anomalies, the spatial location and cross-sectional area of ​​each anomaly, and its estimated length in the tunnel excavation direction are extracted, and the water inrush risk index of each karst cave section is calculated in a comprehensive manner.

[0053] Please see Figure 2 As shown, the calculation process of the water inrush risk index includes: comparing the number of anomalies with the area of ​​the anomaly region to obtain the anomaly distribution density, and comparing the sum of the cross-sectional areas of all anomalies with the area of ​​the anomaly region to obtain the anomaly coverage density.

[0054] The concentrated length and maximum length of each anomaly are calculated from the estimated length of each anomaly in its tunnel excavation direction, and the degree of anomaly of the anomaly is calculated based on the concentrated length and maximum length.

[0055] Based on the spatial location of each anomaly, the vertical distance between the center of each anomaly and the tunnel excavation face is calculated, the shortest vertical distance is extracted, and the relative deviation between the shortest vertical distance and the preset risk distance is calculated to obtain the anomaly distance risk level.

[0056] The distribution density, coverage density, degree of anomaly, and distance risk of anomalies were subjected to minimum-maximum normalization, and the water inrush risk index of each karst cave section was obtained by weighted summation.

[0057] It should be noted that the weights of the above-mentioned anomaly distribution density, anomaly coverage density, anomaly degree, and anomaly distance risk were obtained through industry experience.

[0058] Taking a long, deep-buried tunnel as an example, due to its great depth and high water head pressure, the degree of anomaly of the anomaly directly determines the intensity of water inrush, while the distance risk affects the suddenness of water inrush. The weight of the two is higher. The example weight of the degree of anomaly is 0.35, the example weight of the distance risk of the anomaly is 0.3, the example weight of the anomaly coverage density is 0.2, and the example weight of the anomaly distribution density is 0.15.

[0059] Taking shallow-buried tunnels as an example, due to their low water head pressure, but strong karst development leading to dense distribution of anomalies, the coverage density and distribution density of anomalies directly affect the probability of water inrush, thus increasing their weight. Therefore, the example weight of the degree of anomaly of anomalies is 0.32, the example weight of the risk degree of anomaly distance is 0.28, the example weight of the coverage density of anomalies is 0.22, and the example weight of the risk degree of anomaly distance is 0.18.

[0060] The effective water storage volume of each anomaly is calculated using a water-rich anomaly volume estimation model based on the cross-sectional area of ​​each anomaly and the estimated length in the tunnel excavation direction.

[0061] It should be added that existing technologies often use the total area of ​​the abnormal region or the volume of the largest abnormal body as the core of risk assessment, ignoring that small and dense groups of abnormal bodies may be more likely to cause multi-point water inrush than a single large abnormal body, and also do not consider the coverage ratio of abnormal bodies in the region. This invention accurately captures the risk difference between dense small anomalies and sparse large anomalies by synergizing distribution density and coverage density, and solves the defect that a single area index cannot distinguish distribution characteristics.

[0062] Furthermore, by calculating the degree of anomaly through the concentrated length and the maximum length, it can not only reflect the concentrated trend of the overall water-rich scale, but also identify the potential threat of extreme anomalies. This avoids exaggerating or underestimating the overall risk due to the single maximum length indicator. The distance risk degree is obtained by calculating the relative deviation value between the shortest vertical distance and the preset risk distance, which deeply binds the risk assessment with the construction progress. As the tunnel is excavated, the distance between the anomaly and the excavation face shortens, and the distance risk degree dynamically increases, which can trigger early warning and make the dynamic assessment adapt to the dynamic construction.

[0063] Specifically, the calculation process of the effective water storage volume includes: comparing the relative amplitude value of each sampling point in the reflection anomaly area with the average relative amplitude value of all sampling points in the karst cave section where the reflection anomaly area is located to obtain the amplitude value difference ratio.

[0064] The ratio of the number of sampling points with relative amplitude values ​​greater than a preset amplitude threshold to the total number of sampling points in the abnormal reflection area is used as the local amplitude ratio.

[0065] Extract the mean of the relative amplitude values ​​of the sampling points within the reflection anomaly area, and statistically analyze the mean of the relative amplitude values ​​of all sampling points within the karst cave section where the reflection anomaly area is located. Compare the two to obtain the amplitude difference ratio.

[0066] Similarly, the waveform continuity index of the reflection anomaly region is calculated using the same method as that used for the radar detection region. The waveform continuity index of the reflection anomaly region is then compared with that of the radar detection region to obtain the waveform continuity difference ratio.

[0067] The reflection anomaly factor is obtained by weighted summing the amplitude difference ratio, local amplitude ratio, amplitude difference ratio, and waveform continuity difference ratio in the reflection anomaly region. .

[0068] Match the corresponding baseline porosity based on the geological description of the karst cave and surrounding rock. .

[0069] The formula for calculating the effective water storage volume is as follows: .

[0070] In the formula, This represents the effective water storage volume of the anomalous body. is the cross-sectional area of ​​the anomalous body. This represents the estimated length of the anomaly in the tunnel excavation direction.

[0071] It should be added that the weights of the amplitude value difference ratio, local amplitude ratio, amplitude difference ratio and waveform continuity difference ratio of the above-mentioned reflection anomaly regions are obtained by collecting data on the above four ratio parameters corresponding to multiple reflection anomaly regions, and calculating the information entropy of each ratio parameter. Based on the principle that the smaller the entropy value, the higher the weight, the complement of the information entropy of each indicator is used as the information utility value of each indicator, and the ratio of the information utility value to the sum of all information utility values ​​is used as the weight of the corresponding indicator.

[0072] It should be noted that the above-mentioned benchmark porosity is obtained by comprehensively considering geological characteristics, industry standards, and the specific scenario of the tunnel and karst cave. Taking a complete, unfilled limestone karst cave as an example, when the tunnel and karst cave is in the middle development stage and the groundwater activity is moderate, the benchmark porosity is 0.02. Taking a broken, semi-filled dolomite karst cave as an example, when the tunnel and karst cave is in the late development stage and the groundwater activity is active, the benchmark porosity is 0.052.

[0073] The estimated water inflow of each karst cave section is calculated based on the water inflow risk index and effective water storage volume of each section. The estimated water inflow of the tunnel karst cave is then comprehensively predicted according to the preset prediction rules and output.

[0074] Specifically, the calculation process for the estimated water inflow includes: extracting the maximum height from the height of each karst cave section, and comparing the height of each karst cave section with the maximum height to obtain a height correction coefficient.

[0075] The dynamically corrected inrush risk index is obtained by multiplying the height correction factor by the inrush risk index.

[0076] The effective water storage volume of all anomalies in each karst cave section is summed to obtain the total effective water storage volume of the corresponding karst cave section. This total effective water storage volume is then multiplied by the corresponding corrected water inrush risk index and output as the final total effective water storage volume.

[0077] The types of each cave segment are identified based on the geological description of the cave and the surrounding rock. These types include independent and non-independent types.

[0078] If a certain karst cave section is an independent type, retrieve the release time that matches the corresponding geological description, divide the final total effective water storage volume by the matched release time, and output the estimated water inflow.

[0079] If a certain karst cave section is not independent, retrieve the preset safe release time threshold and perform the same calculation as the estimated inflow of water for independent types.

[0080] In one specific embodiment, the geological description mainly includes the filling situation, rock mass structure and scale, etc.

[0081] Among them, the filling condition is such as no filling and full filling with mud and sand, the rock mass structure is such as intact rock mass, fragmented rock mass, dissolution fissures, etc., and the scale is such as small solution holes and large solution cavities, etc.

[0082] The matching of geological descriptions and release times is achieved through a preset correspondence: for example, when the geological description indicates complete filling with silt and sand, a longer release time, such as 10 hours, is matched because the water body is impeded by the filling material. When the description indicates unfilled large cavities, a very short release time, such as 0.5 hours, is matched because the water body can collapse instantly. If the description indicates partially filled network fractures, an intermediate value, such as 2 hours, is taken, thus achieving precise time matching based on geological characteristics.

[0083] It should be noted that the above-mentioned safe release time threshold is a preset safe release time threshold for high-risk non-independent karst caves. It is mainly used to simulate the worst instantaneous collapse conditions by adopting an extremely conservative short time. For example, the safe release time threshold is set to 30 minutes.

[0084] Specifically, the process of identifying the type of karst cave section includes: extracting geological type, topography and hydrological signs from the geological description, and combining the anomaly outline and spatial location to determine whether all of the following conditions are met: Condition 1: The tunnel is located in an underground watershed, in a high-lying vadose zone, or surrounded by a regional aquitard.

[0085] Condition 2: There are no large springs, underground river outlets, wetlands, or seasonal rivers closely associated with groundwater.

[0086] Condition 3: The anomaly has clear boundaries, a relatively regular shape, and a limited spatial distribution, and is not connected to other anomalies in the surrounding area.

[0087] If conditions 1, 2, and 3 are met simultaneously, the cave segment is judged to be an independent type; otherwise, it is judged to be a non-independent type.

[0088] Among them, a clear boundary of an anomaly means that in the profile and planar data detected by 3D radar, there is a significant and clear boundary between the reflected signals of the anomaly and the surrounding normal rock medium. Specifically, the quantitative and visual characteristics meet the following conditions: the amplitude and phase of the reflected wave at the boundary of the anomaly show a step change; in the three-dimensional modeling results of the 3D radar volume data, the boundary of the anomaly can form a continuous, closed and non-dispersive line or surface through the difference in reflected signals; the reflected signal within the boundary of the anomaly is limited to the boundary range and does not overlap or merge with the layered reflection of the surrounding normal limestone; engineers can directly mark the boundary range using conventional radar data processing software with an error of less than or equal to 0.3 meters.

[0089] Among them, the relatively regular shape means that the three-dimensional shape of the anomaly is close to a geometric regular shape, without obvious irregular branches, abrupt protrusions or diffuse extensions. Specifically, it meets the following conditions: the shape of the anomaly can be classified into basic geometric shapes such as sphere, ellipsoid, cylinder, and near cube; in different detection profiles of the anomaly, the variation range of its key dimensions is less than or equal to 20%; there are no branched small caves or diffuse fissure networks or other auxiliary structures inside the anomaly, or the volume of the auxiliary structures is less than or equal to 10% of the volume of the main anomaly.

[0090] Among them, the spatial distribution limitation refers to the limited spatial extension range of the anomaly and the lack of connection with other anomalies in the surrounding area. Specifically, it meets the following conditions: the three-dimensional volume of the anomaly is less than or equal to 200 cubic meters, the shortest straight-line distance between the anomaly and other surrounding reflective anomalies is greater than or equal to 3 meters, and there is no continuous reflection signal band connecting the two. The anomaly does not cross the preset karst cave segment boundary of the tunnel, ensuring that the anomaly in a single karst cave segment is independent and controllable.

[0091] All the specific values ​​mentioned above can be dynamically adjusted according to the actual scene where the tunnel is located, and are not fixed values.

[0092] It should be added that, firstly, the tunnel is located in an underground watershed, within a high-lying vadose zone, or surrounded by a regional impermeable layer. This ensures that the area is in the recharge or isolation zone of the groundwater system, lacking continuous recharge from regional groundwater, and the water is mostly stored locally. Secondly, there are no large springs, underground river outlets, wetlands, or other groundwater discharge characteristics, which excludes the existence of strong runoff channels and concentrated discharge points, further confirming the closed nature of the water system. Finally, geophysical exploration shows that the anomaly has clear boundaries, a regular shape, and a limited spatial distribution, with no connection to surrounding anomalies. This indicates that the water storage space is independent and closed, rather than connected to a pipeline network. Only when all three of the above conditions are met can it be determined to be an independent karst cave, with its water inflow mainly coming from static reserves, and the risk is relatively controllable.

[0093] If any of the conditions are not met, especially if discharge characteristics are present, the cave is located in a runoff area, or is connected to an anomaly, it is determined to be a non-independent karst cave because it has a continuous recharge capacity and a high risk of water inrush.

[0094] Please see Figure 3 As shown, the prediction process of the prediction rule includes: calculating the ratio of the number of karst cave segments with a water inrush risk index greater than a preset water inrush risk index threshold to the total number of karst cave segments to obtain the proportion of karst cave segments with water inrush risk.

[0095] If the ratio is greater than the preset ratio threshold, the sum of the estimated water inflow of each karst cave section is directly output as the estimated water inflow of the tunnel karst cave.

[0096] Conversely, the final estimated water inflow is sorted from largest to smallest, and a baseline water inflow is determined based on the sorting.

[0097] The target karst cave segment is the karst cave segment where the benchmark inflow volume is located. The ratio of the sum of the estimated inflow volumes of all other karst cave segments to the total estimated inflow volume of all karst cave segments in the tunnel is used to obtain the proportion of the total estimated inflow volume.

[0098] Similarly, the percentage of effective water storage volume of the total anomaly body is calculated using the same method as the percentage of total estimated inflow.

[0099] Based on the average water inflow risk index of all other karst cave sections, the proportion of total estimated water inflow, and the proportion of effective water storage volume of the total anomaly, a comprehensive correction coefficient is obtained by weighted summation.

[0100] The additional supplementary inflow is obtained by multiplying the comprehensive correction factor by the baseline inflow, and the final estimated inflow of the tunnel karst cave is obtained by adding the additional supplementary inflow to the baseline inflow.

[0101] It should be noted that the above-mentioned preset water inrush risk index threshold is based on an initial threshold obtained from industry experience. Then, it is adjusted according to the tunnel engineering level to obtain the final preset water inrush risk index threshold. Tunnel engineering levels are divided into Level 1, Level 2 and Level 3. The higher the engineering level, the stricter the standards for preventing and controlling water inrush risk, and thus the lower the tolerance for the water inrush risk index. Based on this, the preset water inrush risk index threshold decreases as the engineering level increases, thereby ensuring that high-level projects trigger early warnings at a lower risk index and achieve stricter risk control.

[0102] For example, taking Level 1 as an example, the preset water inrush risk index threshold is 0.21; taking Level 2 as an example, the preset water inrush risk index threshold is 0.33; and taking Level 3 as an example, the preset water inrush risk index threshold is 0.41.

[0103] It should be noted that the aforementioned preset ratio thresholds are obtained based on industry experience. Taking traffic tunnels as an example, when the traffic tunnel is a high-speed rail tunnel or a subway tunnel, the preset ratio threshold is 0.15. When the traffic tunnel is a regular highway tunnel, the preset ratio threshold is 0.25.

[0104] Furthermore, the initial weights of the water inrush risk index, the proportion of total estimated water inrush volume, and the proportion of total effective water storage volume of the anomaly for all other karst cave sections were obtained through industry experience, and then adjusted in conjunction with the tunnel safety level to obtain the final weights of the water inrush risk index, the proportion of total estimated water inrush volume, and the proportion of total effective water storage volume of the anomaly.

[0105] For example, taking a Class I tunnel as an example, the weight of the water inrush risk index for all other karst cave sections is 0.43, the weight of the proportion of the total estimated water inrush volume for all other karst cave sections is 0.37, and the weight of the proportion of the total effective water storage volume of the total anomalous body for all other karst cave sections is 0.20. Taking a Class II tunnel as an example, the weight of the water inrush risk index for all other karst cave sections is 0.38, the weight of the proportion of the total estimated water inrush volume for all other karst cave sections is 0.40, and the weight of the proportion of the total effective water storage volume of the total anomalous body for all other karst cave sections is 0.22.

[0106] Please see Figure 4 As shown, the process of determining the benchmark inflow includes: if the height of the first-ranked karst cave segment is not the lowest height, then the estimated inflow of the first-ranked karst cave segment is used as the benchmark inflow.

[0107] If it is the lowest height, then extract the second-ranked karst cave segment and determine whether the height of the karst cave segment is the lowest height. If not, then select the estimated inflow volume of the second-ranked segment as the benchmark inflow volume.

[0108] If the height of the selected karst cave segment is the lowest height, the lowest height is iterated and judged in the sorting order until the height of the selected karst cave segment is not the lowest height, and the estimated water inflow of the corresponding selected karst cave segment is used as the benchmark water inflow.

[0109] It should be added that, since existing technologies often directly take the maximum inflow segment as the benchmark, they are prone to mistakenly taking the lowest static water accumulation segment as the core risk source, resulting in deviation in the direction of prevention and control. Therefore, it is necessary to divide the scenarios by setting a preset proportion threshold. When the proportion exceeds the preset proportion threshold, prevention and control should be coordinated according to the total inflow. When the proportion is lower than the preset proportion threshold, the focus should be on the key high inflow segment. This will enable differentiated decision-making for coordinated prevention of high-risk segments and precise prevention of low-risk segments, balancing safety and cost.

[0110] Furthermore, during tunnel construction, the water inflow in the lowest elevation section may be relatively large, but it is mostly static water accumulation. In contrast, the side wall karst caves near the tunnel face are more prone to dynamic water inflow due to construction disturbances. By iteratively judging and excluding the lowest elevation section, and prioritizing the high water inflow section at the non-lowest elevation as the benchmark water inflow, we can accurately identify the dynamic risk sources that have a greater impact on construction.

[0111] Meanwhile, a comprehensive correction coefficient is calculated by weighting the inrush risk index of the unselected section, the proportion of the total estimated inrush volume, and the proportion of the effective water storage volume of the total anomaly. Each dimension comes from quantitative data, which enables the additional inrush volume to accurately match the actual risk of the unselected section, reduce prediction errors, and make the calculation results more in line with the actual project. Each step is based on objective data and the logic is traceable, which improves the credibility of the prediction results and the engineering guidance value.

[0112] Please see Figure 5 As shown, the present invention also provides a tunnel karst cave water inflow prediction and analysis system based on ground-penetrating radar. The system includes: a karst cave division module, an abnormal area identification module, a water inflow risk calculation module, a water storage information acquisition module, and a water inflow prediction module.

[0113] In the above, the abnormal area identification module is connected to the karst cave division module and the water inrush risk calculation module, respectively. The water storage information acquisition module is connected to the water inrush risk calculation module and the water inrush volume prediction module, respectively. The water inrush volume prediction module is also connected to the water inrush risk calculation module.

[0114] The cave segmentation module divides the tunnel caves into sections according to a preset height, and uses 3D radar to detect each cave section and mark the height of each cave section.

[0115] The abnormal region identification module obtains the amplitude intensity, hyperbolic shape and disorder level based on the detection data, and identifies abnormal reflection regions.

[0116] The water inrush risk calculation module extracts the area of ​​the abnormal area, the number of abnormal bodies, the spatial location and cross-sectional area of ​​each abnormal body, and its estimated length in the tunnel excavation direction for the karst cave section with anomaly reflection areas, and calculates the water inrush risk index of each karst cave section in a comprehensive manner.

[0117] The water storage information acquisition module calculates the effective water storage volume based on the cross-sectional area of ​​each anomaly and the estimated length in the tunnel excavation direction using a water-rich anomaly volume estimation model.

[0118] The water inflow prediction module comprehensively predicts the estimated water inflow of the tunnel cave based on the water inflow risk index and the effective water storage volume, and outputs the result.

[0119] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A tunnel karst cave water inflow prediction analysis method based on ground penetrating radar, characterized in that, The method comprises: The tunnel cave is divided into several cave sections according to the preset height from top to bottom, and each cave section is detected by a 3D radar to obtain 3D radar volume data, and the height of each cave section is labeled; According to the 3D radar volume data, the amplitude intensity, hyperbolic shape and disorder degree are obtained to identify the reflection anomaly area; For the cave section with the reflection anomaly area, the area of the abnormal area, the number of abnormal bodies, the spatial position of each abnormal body, the cross-sectional area and the estimated length in the tunnel excavation direction are extracted, and the water gushing risk index of each cave section is calculated; According to the cross-sectional area and the estimated length of each abnormal body in the tunnel excavation direction, the effective water storage volume of each abnormal body is calculated through the water-rich abnormal body volume estimation model; According to the water gushing risk index and the effective water storage volume of each cave section, the estimated water gushing amount of each cave section is calculated, and the estimated water gushing amount of the tunnel cave is comprehensively predicted according to the preset prediction rule and output.

2. The ground penetrating radar-based tunnel cave-inflow prediction analysis method of claim 1, wherein: The amplitude intensity acquisition process comprises: The amplitude of each sampling point in the corresponding depth range of each cave section is extracted from the reflection wave signal waveform, and the relative amplitude value is obtained after minimum-maximum normalization processing; The average value of the relative amplitude value of each sampling point is calculated, and the maximum relative amplitude value is selected; The ratio of the number of sampling points greater than the preset amplitude threshold to the total number of sampling points in each cave section is calculated as the amplitude ratio; The amplitude intensity is output by comparing the amplitude ratio with the preset amplitude ratio threshold. 3.The geological radar-based tunnel cave water inflow prediction analysis method of claim 1, wherein: The disorder degree acquisition process comprises: The waveform profile of the radar detection area is intercepted in the excavation direction, the transverse direction and the height direction, and the amplitude value and the phase of each sampling point in the radar detection area are extracted; The amplitude difference of all adjacent sampling points is calculated, and the average amplitude difference of adjacent sampling points is calculated, and the maximum amplitude value is extracted; The phase difference average value of adjacent sampling points is calculated in the same way as the amplitude difference average value; The waveform continuity index of the radar detection area is calculated based on the amplitude difference average value, the maximum amplitude value and the phase difference average value, and the complement of the index is taken as the disorder degree. 4.The geological radar-based tunnel cave water inflow prediction analysis method of claim 1, wherein: The water gushing risk index calculation process comprises: The abnormal body distribution density is obtained by comparing the number of abnormal bodies with the area of the abnormal area, and the abnormal body coverage density is obtained by comparing the sum of the cross-sectional areas of each abnormal body with the area of the abnormal area; The concentrated length and the maximum length are counted from the estimated length of each abnormal body in the tunnel excavation direction, and the abnormal degree of the abnormal body is counted based on the concentrated length and the maximum length; Based on the spatial position of each abnormal body, the vertical distance between the center of each abnormal body and the tunnel excavation surface is calculated, the shortest vertical distance is extracted, and the relative deviation value of the shortest vertical distance and the preset risk distance is calculated to obtain the abnormal body distance risk degree; The abnormal body distribution density, the abnormal body coverage density, the abnormal body abnormal degree and the abnormal body distance risk degree are subjected to minimum-maximum normalization processing, and the water gushing risk index of each cave section is calculated by weighted summation. 5.The geological radar-based tunnel cave water inflow prediction analysis method of claim 2, wherein: The effective water storage volume calculation process comprises: The amplitude value difference ratio is obtained by comparing the relative amplitude value of each sampling point in the reflection anomaly area with the average relative amplitude value of all sampling points in the cave section where the reflection anomaly area is located; The ratio of the number of sampling points with a relative amplitude value greater than a preset amplitude threshold to the total number of sampling points in the reflection anomaly region is taken as a local amplitude ratio; The mean value of the relative amplitude values of the sampling points in the reflection anomaly region is extracted, and the mean value of the relative amplitude values in the cave segment where the reflection anomaly region is located is counted, and the ratio of the two is taken as an amplitude difference ratio; The waveform continuity index of the reflection anomaly region is calculated in the same way as the waveform continuity index of the radar detection region, and the ratio of the waveform continuity indexes of the reflection anomaly region and the radar detection region is taken as a waveform continuity difference ratio; The amplitude value difference proportion, local amplitude proportion, amplitude difference proportion and waveform continuity difference proportion of the reflection anomaly area are weighted and summed to obtain a reflection anomaly factor ; Matching the corresponding baseline porosity according to the geological description of the cave and the surrounding rock ; The effective water storage volume of the abnormal body is calculated by the following formula ; wherein, is the effective water storage volume of the abnormal body; is the cross-sectional area of the abnormal body; is the presumed length of the abnormal body in the tunneling direction. 6.The geological radar-based tunnel cave water inflow prediction analysis method according to claim 1, wherein: The calculation process of the estimated water inflow includes: The maximum height is extracted from the heights of the cave segments, and the height of each cave segment is divided by the maximum height to obtain a height correction coefficient; The height correction coefficient is multiplied by the water inflow risk index to obtain a dynamically corrected water inflow risk index; The effective water storage volumes of all abnormal bodies in each cave segment are summed to obtain the total effective water storage volume of the corresponding cave segment, and the total effective water storage volume is multiplied by the corresponding corrected water inflow risk index to output the final total effective water storage volume; The types of the cave segments are identified according to the geological description of the caves and surrounding rocks, and the types include independent type and non-independent type; If a cave segment is of the independent type, the release time matched with the corresponding geological description is called, and the final total effective water storage volume is divided by the matched release time to output the estimated water inflow; If a cave segment is of the non-independent type, a preset safe release time threshold is called, and the estimated water inflow is calculated in the same way as the estimated water inflow of the independent type.

7. The ground penetrating radar-based tunnel cave-inflow prediction analysis method of claim 6, wherein: The identification process of the cave segment types includes: The geological types, topography and hydrological signs are extracted from the geological description, and the abnormal body profile and spatial position are combined to determine whether all the following conditions are met: Condition 1: The tunnel is located in the underground divide, the high-relief aeration zone, or surrounded by regional aquifuge; Condition 2: There is no large spring, underground river outlet, wetland or seasonal river closely related to groundwater; Condition 3: The abnormal body boundary is clear, the shape is relatively regular, and the spatial distribution is limited, and there is no connection with other abnormalities around; If conditions 1, 2 and 3 are met at the same time, the cave segment is determined to be of the independent type, otherwise it is determined to be of the non-independent type. 8.The geological radar-based tunnel cave water inflow prediction analysis method of claim 1, wherein: The prediction process of the prediction rule includes: The ratio of the number of cave segments with a water inflow risk index greater than a preset water inflow risk index threshold to the total number of cave segments is taken as a water inflow risk cave segment ratio; If the ratio is greater than a preset proportion threshold, the sum of the estimated water inflow of each cave segment is directly output as the estimated water inflow of the tunnel cave; Otherwise, the final estimated water inflow is sorted from large to small, and a reference water inflow is determined based on the sorting; The cave segment where the reference water inflow is located is taken as the target cave segment, the sum of the estimated water inflows of all other cave segments is divided by the total estimated water inflow of the tunnel cave to obtain a total estimated water inflow proportion; The total abnormal body effective water storage volume proportion is calculated in the same way as the total estimated water inflow proportion; Based on the mean value of the water inflow risk indexes of all other cave segments, the total estimated water inflow proportion and the total abnormal body effective water storage volume proportion, a comprehensive correction coefficient is obtained by weighted summation; The additional supplementary gushing water is obtained by multiplying the comprehensive correction coefficient and the reference gushing water, and the final estimated gushing water of the tunnel karst cave is obtained by adding the additional supplementary gushing water and the reference gushing water. 9.The tunnel cave water inflow prediction analysis method based on ground penetrating radar according to claim 8, wherein: The determination process of the reference gushing water comprises: If the height of the karst cave segment ranked first is not the lowest height, the estimated gushing water of the karst cave segment ranked first is taken as the reference gushing water; If the height of the karst cave segment ranked first is the lowest height, the karst cave segment ranked second is extracted, and whether the height of the karst cave segment is the lowest height is judged, if not, the estimated gushing water of the karst cave segment ranked second is taken as the reference gushing water; If the height of the karst cave segment is the lowest height, the lowest height iterative judgment is carried out according to the ranking order, until the height of the selected karst cave segment is not the lowest height, and the estimated gushing water of the corresponding selected karst cave segment is taken as the reference gushing water.

10. A geological radar-based tunnel karst cave water inflow prediction analysis system, characterized in that: The system comprises: The karst cave division module segments the tunnel karst cave according to the preset height, detects each karst cave segment by the 3D radar, obtains the 3D radar volume data, and labels the height of the karst cave segment; The abnormal area identification module obtains the amplitude intensity, hyperbolic shape and disorder degree according to the 3D radar volume data, and identifies the reflection abnormal area; The gushing risk calculation module extracts the abnormal area area, the number of abnormal bodies, the spatial position of each abnormal body, the cross-sectional area and the estimated length of each abnormal body in the tunnel excavation direction of the karst cave segment with the reflection abnormal area, and comprehensively calculates the gushing risk index of each karst cave segment; The water storage information acquisition module calculates the effective water storage volume by the water-rich abnormal body volume estimation model according to the cross-sectional area of each abnormal body and the estimated length in the tunnel excavation direction; The gushing amount prediction module comprehensively predicts the estimated gushing water of the tunnel karst cave according to the gushing risk index and the effective water storage volume, and outputs.

Images