Methods and apparatus for analyzing adverse geological problems in tunnels using multi-source data fusion

Abstract

This application relates to the field of geological analysis technology and discloses a method and apparatus for analyzing adverse geological problems in tunnels using multi-source data fusion. It aims to address the problem of poor accuracy in existing analysis schemes for adverse geological problems in tunnels. The scheme mainly includes: establishing a full-process information database for tunnels and an index system for adverse geological problems; determining the detection indicators and corresponding detection data for the adverse geological problems to be analyzed; converting the detection data into risk values; dividing the detection interval into multiple segments according to the coverage of different detection data; and determining the weight of each detection indicator; calculating the average risk index of each segment based on the risk value and weight of each detection indicator, using methods such as weighted average, fuzzy fusion theory, Laplace fusion theory, Bayesian averaging, and robust statistics; and determining whether the adverse geological problem to be analyzed will occur in each segment based on the average risk index. This application improves the accuracy of analyzing adverse geological problems in tunnels.

Description

Technical Field

[0001] This application relates to the field of geological analysis technology, specifically to a method and apparatus for analyzing adverse geological problems in tunnels using multi-source data fusion. Background Technology

[0002] Tunnel advance geological prediction is a method of multi-scale, multi-parameter forecasting of the geological conditions ahead of the working face, based on site conditions and employing one or more methods such as geological analysis, advance geological drilling, geophysical exploration, and advance pilot tunneling. Through advance geological prediction, adverse geological conditions and major geological problems ahead of the tunnel can be identified, and potential adverse geological problems such as rock bursts and collapses can be analyzed, providing early warnings for construction, ensuring construction safety, and providing a reliable basis for information-based design and construction.

[0003] Existing schemes for analyzing adverse geological problems using tunnel advanced geological prediction have at least the following problems: First, existing tunnel advanced geological prediction methods, when conducting comprehensive exploration, mostly rely primarily on geophysical information, supplemented by geological information. In actual engineering, while methods such as geophysical exploration, sensor detection, and simulation experiments have advantages such as non-destructive exploration and high efficiency, their comparison with geological sketches and advanced drilling results is not intuitive enough. Moreover, when there is significant interference from tunnel environmental factors, complex regional geological conditions, and frequent adverse geological problems, the accuracy of the analysis of adverse geological conditions is prone to large deviations. Second, existing multi-source data fusion methods consider limited information sources and do not fully consider the input of information from the entire tunnel excavation process, thus reducing the accuracy of the analysis. Third, due to unclear geological logic in the interpretation, existing intelligent machine interpretation of advanced geological prediction still relies mainly on manual interpretation, resulting in poor accuracy and universality. Summary of the Invention

[0004] This application aims to address the problem of poor accuracy in existing analytical schemes for adverse geological problems in tunnels, and proposes a multi-source data fusion method and apparatus for analyzing adverse geological problems in tunnels.

[0005] The technical solution adopted by this application to solve the above-mentioned technical problems is:

[0006] Firstly, this application provides a method for analyzing adverse geological problems in tunnels using multi-source data fusion, the method comprising:

[0007] Establish a comprehensive information database for tunnel construction and an indicator system for adverse geological problems;

[0008] Identify the current unfavorable geological problems to be analyzed in the tunnel, determine the detection indicators corresponding to the unfavorable geological problems to be analyzed in the unfavorable geological problem index system, and determine the detection data of each detection indicator within the detection interval in the whole process information database;

[0009] The detection data of each detection indicator is converted into risk values. The detection interval is divided into multiple segments according to the coverage of different detection data in each detection indicator, and the weight of each detection indicator is determined.

[0010] Based on the risk value and weight corresponding to each detection indicator, and using the weighted average method, fuzzy fusion theory, Laplace fusion theory, Bayesian averaging method, and robust statistical method, the average risk index of each section is calculated. Based on the average risk index, it is determined whether the undesirable geological problems to be analyzed will occur in each section.

[0011] Furthermore, the entire process information database includes basic tunnel design information, basic geological condition information, advanced exploration information, advanced geophysical exploration information, excavation and construction parameter information, test and inspection information, and monitoring and detection information;

[0012] The basic design information includes tunnel morphology, span, height, excavation method, and surrounding rock grade; the basic geological condition information includes geomorphological features, vertical burial depth, stratigraphic code, lithology, saturation coefficient, rock mass integrity, groundwater pressure, and shear strength; the advanced exploration information includes backflow volume, backflow color, and karst cave conditions corresponding to advanced boreholes, advanced exploratory tunnels, and advanced exploratory wells; the advanced geophysical exploration information includes information corresponding to seismic wave methods, ground-penetrating radar methods, transient electromagnetic methods, and induced polarization methods; the excavation and construction parameter information includes information corresponding to drill-and-blast method construction records, TBM method construction records, and rock debris records; the test and testing information includes information corresponding to rock tests and tests, soil tests and tests, and water tests and tests; and the monitoring and detection information includes information corresponding to microseismic monitoring, deformation monitoring, and radioactivity detection.

[0013] Furthermore, the adverse geological problems to be analyzed include high-stress rockbursts, high-stress large deformations, karst caves, karst water and mud inrushes, fault fracture zone collapses, fault fracture zone water and mud inrushes, large deformations in weak rock zones, collapses in weak rock zones, high geothermal heat hazards, toxic gas hazards, and radioactive hazards.

[0014] Furthermore, the detection data of each indicator are converted into risk values, specifically including:

[0015] The detection index with upper and lower limits and continuous intermediate values ​​is used as the first detection index. The value range of the first detection index is divided into multiple intervals, and a first mapping relationship between each interval and the risk value is established.

[0016] A first mapping function is established between the first detection indicator and the risk value based on the first mapping relationship. The corresponding risk value is determined based on the value corresponding to the detection data of the first detection indicator and the first mapping function.

[0017] Furthermore, the detection data of each indicator are converted into risk values, specifically including:

[0018] The detection index with interval description but no specific value is used as the second detection index. The second mapping relationship between each interval description of the second detection index and the risk value is established. In the second mapping relationship, the interval description is added and the risk value is interpolated accordingly.

[0019] The corresponding risk value is determined based on the description of the detection data of the second detection indicator and the second mapping relationship.

[0020] Furthermore, the detection data of each indicator are converted into risk values, specifically including:

[0021] The detection index with only qualitative description is used as the third detection index. A third mapping relationship is established between each qualitative description of the third detection index and the risk value. The corresponding risk value is determined based on the qualitative description corresponding to the detection data of the third detection index and the third mapping relationship.

[0022] Furthermore, based on the coverage range of different detection data in each detection indicator, the detection interval is divided into multiple segments, specifically including:

[0023] The coverage range of different detection data for each detection indicator in the tunnel is determined. There are endpoints between the coverage ranges of different detection data. The location of each endpoint of each detection indicator is used as a segmentation point. The detection interval is divided into multiple segments according to the segmentation points.

[0024] Furthermore, the weights of each detection indicator are determined by using the analytic hierarchy process (AHP) to calculate the first weight of each detection indicator, and using the Huber weighting method to calculate the second weight of each detection indicator. The average of the first and second weights is then used as the weight of each detection indicator.

[0025] Furthermore, the method for calculating the average risk index includes:

[0026] The first risk index of each segment is calculated based on the weighted average method, the second risk index of each segment is calculated based on the fuzzy fusion theory, the third risk index of each segment is calculated based on the Laplace fusion theory, the fourth risk index of each segment is calculated based on the Bayesian average method, and the fifth risk index of each segment is calculated based on the robust statistical method.

[0027] The average risk index is obtained by calculating the average of the first, second, third, fourth, and fifth risk indices.

[0028] Secondly, this application provides a multi-source data fusion analysis device for tunnel adverse geological problems, the device being used to implement the steps of the multi-source data fusion analysis method for tunnel adverse geological problems as described in the first aspect.

[0029] The beneficial effects of this application are as follows: The method and apparatus for analyzing adverse geological problems in tunnels using multi-source data fusion provided in this application take geological information as the core, constructs an information database of the entire tunnel excavation process, and establishes a set of indicator systems for adverse geological problems in tunnel engineering. When it is necessary to analyze adverse geological problems, the relevant detection indicators are determined from the indicator system, and the detection data corresponding to each detection indicator is obtained from the entire process information database. Based on the obtained detection data, risk value conversion and alignment processing are performed. Then, information technology is introduced, and mathematical geological theory is applied to carry out matrix fusion calculation of multi-source data, realizing comprehensive analysis of multi-source data fusion, obtaining the analysis results of adverse geological problems in each section, realizing the optimization of complementary and redundant information in space and time within the indicator system, obtaining a consistent interpretation of the predicted object, and improving the accuracy of adverse geological problem analysis. Attached Figure Description

[0030] Figure 1 A flowchart illustrating a method for analyzing adverse geological problems in tunnels using multi-source data fusion, provided in an embodiment of this application;

[0031] Figure 2 This is a schematic diagram illustrating the division of a detection interval into multiple segments, as provided in an embodiment of this application. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0033] In some of the processes described in the specification and accompanying drawings of this application, multiple operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The sequence numbers of the operations are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0034] The technical solutions of this application are applicable to application scenarios that require analysis of whether tunnels will experience adverse geological problems, such as analyzing whether tunnels will experience geological disasters such as water inrush, gas outburst, rock burst, or large deformation.

[0035] Current methods for analyzing adverse geological problems in tunnels typically rely on advanced geological prediction for detection. However, existing advanced geological prediction methods consider limited information sources and do not adequately account for the input of information throughout the entire tunnel excavation process, which reduces the accuracy of the analysis. Furthermore, the accuracy of the analysis is affected by factors such as the tunnel environment, the complexity of regional geological conditions, and the frequency of adverse geological problems. In addition, the presentation of analysis results is usually based on manual interpretation, which further reduces the accuracy of the analysis results.

[0036] Based on this, the technical solution of this application is proposed. In the embodiments of this application, a tunnel full-process information database and an adverse geological problem index system are established; the current adverse geological problems to be analyzed in the tunnel are determined, and the detection indicators corresponding to the adverse geological problems to be analyzed are determined in the adverse geological problem index system; the detection data of each detection indicator within the detection interval are determined in the full-process information database; the detection data of each detection indicator are converted into risk values; the detection interval is divided into multiple segments according to the coverage of different detection data in each detection indicator, and the weight of each detection indicator is determined; the average risk index of each segment is calculated based on the risk value and weight corresponding to each detection indicator and based on the weighted average method, fuzzy fusion theory, Laplace fusion theory, Bayesian averaging method and robust statistical method; and the average risk index is used to determine whether the adverse geological problems to be analyzed will occur in each segment.

[0037] Specifically, this application embodiment takes geological information as the core, establishes a full-process information database using information data from the entire tunnel excavation process, and establishes an indicator system for adverse geological problems based on various adverse geological problems and their related detection indicators. When it is necessary to analyze an adverse geological problem, the relevant detection indicators are first determined according to the adverse geological problem indicator system, and the detection data of each detection indicator are determined based on the full-process information data. Then, risk value conversion, segment division, and weight calculation are performed on each detection indicator. Finally, matrix fusion calculation of multi-source data is carried out based on the risk value and weight of each detection indicator corresponding to each segment to obtain the average risk index of the adverse geological problem to be analyzed in each segment, realizing comprehensive analysis of multi-source data fusion, thereby improving the accuracy of adverse geological problem analysis.

[0038] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0039] Figure 1 A flowchart illustrating a method for analyzing adverse geological problems in tunnels using multi-source data fusion, as provided in this application embodiment, is available for reference. Figure 1 The method includes the following steps:

[0040] S101. Establish a full-process information database for tunnels and an indicator system for adverse geological problems.

[0041] In this embodiment of the application, the whole process information database includes basic tunnel design information, basic geological condition information, advanced exploration (holes, tunnels, wells) information, advanced geophysical exploration information, excavation and construction parameter information, test and inspection information, and monitoring and detection information.

[0042] This application's embodiments extract several key detection indicators for analyzing adverse geological problems, realizing the transformation from scattered information to full-process information, and from fuzzy factors to clear indicators. Specifically, in the basic geological condition information, five subcategories of information are selected: topography, stratigraphy and lithology, geological structure, groundwater, and working face evaluation, including indicators such as geomorphological features, vertical burial depth, stratigraphic code, lithology, saturation coefficient, rock mass integrity, groundwater pressure, and shear strength; in the advanced exploration (borehole, cave, well) information, three subcategories of information are selected: advanced borehole, advanced cave, and advanced well, including indicators such as backflow volume, backflow color, and karst cave conditions; in the advanced geophysical exploration information, subcategories of information are selected, including seismic wave method, ground-penetrating radar method, transient electromagnetic method, and induced polarization method, including... The data includes indicators such as longitudinal wave velocity, Poisson's ratio, resistivity, and polarizability; excavation and construction parameters, including three subcategories: drill-and-blast method construction records, TBM (tunnel boring machine) method construction records, and rock debris records, encompassing indicators such as excavation footage, cutterhead thrust, and rock debris size; testing and experimental information, including three subcategories: rock testing and experimental information, soil testing and experimental information, and water testing and experimental information, encompassing indicators such as the main mineral components of rocks and descriptions of rock grinding discs; and monitoring and detection information, including microseismic monitoring, deformation monitoring, and radioactivity detection, encompassing indicators such as moment magnitude, settlement, and radon concentration.

[0043] In this embodiment, adverse geological problems include high-stress rockbursts, large-scale deformations due to high stress, karst caves, karst water and mud inrushes, fault fracture zone collapses, large-scale deformations in weak rock zones, collapses in weak rock zones, high-temperature heat hazards, toxic gas hazards, and radioactive hazards. In establishing the indicator system for adverse geological problems, various adverse geological problems are analyzed, and corresponding detection indicators are selected from a comprehensive information database to determine the relevant detection indicators for each type of adverse geological problem.

[0044] For example, regarding the adverse geological problem of fault fracture zone collapse, an analysis of fault fracture zone collapse is conducted, and multiple detection indicators related to it are identified by screening from a comprehensive information database. Specifically, a collapse refers to a sudden, destructive geological hazard such as collapse, pile-up, or landslide caused by the instability of the surrounding rock. Before tunnel excavation, the rock mass is in a certain stress equilibrium state. Engineering excavation will cause stress redistribution in the rock mass surrounding the tunnel. When the tunnel passes through areas with low surrounding rock strength and poor stability, such as fault fracture zones and weak rock zones, the excavated surrounding rock cannot adapt to the changes in redistributed stress, and will deform and become unstable inward, ultimately causing a collapse. This embodiment focuses on collapses in fault fracture zones, primarily selecting information such as topography, lithology, geological structure, groundwater, tunnel face evaluation, advanced drilling, seismic wave method, induced polarization method, TBM method construction records, on-site water tests and measurements, deformation monitoring, stress monitoring, and seepage detection. It also includes detection indicators such as tunnel morphology, rotational speed, longitudinal wave velocity, resistivity, settlement, support internal force, and flow rate.

[0045] S102. Determine the current unfavorable geological problems to be analyzed in the tunnel, and determine the detection indicators corresponding to the unfavorable geological problems to be analyzed in the unfavorable geological problem index system, and determine the detection data of each detection indicator within the detection interval in the whole process information database.

[0046] In practical applications, the specific adverse geological problem to be analyzed is determined based on the specific working conditions. Relevant detection indicators are then identified within the adverse geological problem indicator system, and the detection data for each indicator within the detection interval are determined in the full-process information database. For example, if the adverse geological problem to be analyzed is a fault fracture zone collapse, then detection indicators related to fault fracture zone collapse are determined within the adverse geological problem indicator system, and the detection data for each indicator within the detection interval are determined in the full-process information database.

[0047] S103. Convert the detection data of each detection indicator into risk values, divide the detection interval into multiple segments according to the coverage of different detection data in each detection indicator, and determine the weight of each detection indicator.

[0048] In this embodiment, a mapping transformation method is used to convert the detection data into a normalized dimensionless risk value ranging from 0 to 1, which characterizes the probability of occurrence of adverse geological problems under the influence of a single indicator from low to high, and achieves consistency in the quantitative form and data range of each detection indicator. The mapping transformation method includes a piecewise function mapping transformation method for quantitative indicators, a node interpolation mapping transformation method for semi-qualitative indicators, and a threshold-level mapping transformation method for qualitative indicators.

[0049] In this embodiment of the application, the quantitative indicator piecewise function mapping transformation method includes: taking a detection indicator with upper and lower limit values ​​and continuous intermediate values ​​as a first detection indicator; dividing the numerical range of the first detection indicator into multiple intervals and establishing a first mapping relationship between each interval and the risk value; establishing a first mapping function between the first detection indicator and the risk value based on the first mapping relationship; and determining the corresponding risk value based on the numerical value corresponding to the detection data of the first detection indicator and the first mapping function.

[0050] Specifically, for primary detection indicators such as Poisson's ratio and P-wave velocity, which have clearly defined upper and lower limits and continuous intermediate values, the risk value of the corresponding detection data is determined by mapping segmentation, establishing a mapping function, and determining the values ​​of constant parameters. The specific steps are as follows:

[0051] (1) Establish the upper and lower limits of the first detection indicator and its correlation with the risk value. Taking Poisson's ratio as an example, its upper and lower limits range from 0 to 0.5, and it is positively correlated with the risk value (0 to 1). Therefore, we can assume that when the Poisson's ratio is 0, the corresponding risk value is 0, and when the Poisson's ratio is 0.5, the corresponding risk value is 1.

[0052] (2) Mapping and segmenting the first detection index. Based on different strata conditions, specificity and segmentation descriptions are performed. For example, the Poisson's ratio of shale is generally 0.10 to 0.20. We can assume that when the Poisson's ratio is 0.10, the corresponding risk value is 0.2, and when the Poisson's ratio is 0.20, the corresponding risk value is 0.7. Then the Poisson's ratio is divided into three intervals: 0 to 0.10, 0.10 to 0.20, and 0.20 to 0.50. The first mapping relationship with the risk value is established for each interval.

[0053] (3) Establish the first mapping function between the first detection index and the risk value. Generally, three common relational models are selected, namely linear, exponential and S-shaped, but it is not limited to these. The first mapping function can also be other model functions.

[0054] Linear form: y = ax + b;

[0055] Exponential type: y = ae bx +c;

[0056] S-shaped:

[0057] Where y represents the risk value, x represents the detection data corresponding to the first detection indicator, and a, b, and c represent parameters that need to be determined through data fitting. Linear relationships are typically applicable to simple linear relationships, where the risk value increases or decreases linearly with the increase of the first detection indicator. Exponential relationships are applicable to relationships where the first detection indicator and the risk value exhibit exponential growth or decay, and are typically used to describe non-linear growth or decay phenomena. S-shaped relationships are typically used to describe asymptotic growth relationships, where the risk value increases rapidly at first and then tends to plateau with the increase of the first detection indicator.

[0058] (4) Determine the parameter values ​​in the first mapping function. The parameters determine the specific form of the first mapping function and its ability to describe the data. When the amount of data is small, analyze and judge the situation at the working face, refer to the value of the first detection index at the working face and the geological conditions, determine the risk value corresponding to the detection data of the first detection index at the working face, and calculate the parameter value in combination with the upper and lower limits of the interval.

[0059] For more complex first mapping functions, the least squares method can be used for data fitting when the data volume is large, thereby establishing a more realistic first mapping function. The least squares method is the most commonly used basic algorithm; its core idea is to find a curve that minimizes the sum of the squares of the vertical distances from all data points to that curve. Specifically, for n sets of data points (x... i ,y i Given a first mapping function f(x) to be fitted, the least squares method finds the optimal parameters by minimizing the following objective function:

[0060]

[0061] In the formula, S represents the sum of squared errors, and f(x) i ) represents the analytical value of the first mapping function, y i This represents the actual value. By taking the derivative of the parameters and setting the derivative to zero, we can obtain the parameter values ​​that minimize the sum of squared errors. These parameter values ​​minimize the error between the fitted curve and the data points.

[0062] (5) After determining the optimal parameters, the first mapping function can be obtained. Substitute the value corresponding to the detection data of the first detection index into the first mapping function to obtain the corresponding risk value.

[0063] In this embodiment of the application, the semi-qualitative index node interpolation mapping transformation method includes: taking a detection index with an interval description but no specific value as a second detection index, establishing a second mapping relationship between each interval description of the second detection index and the risk value, and adding an interval description and interpolating the risk value accordingly in the second mapping relationship; determining the corresponding risk value based on the description corresponding to the detection data of the second detection index and based on the second mapping relationship.

[0064] Specifically, for semi-qualitative secondary detection indicators such as groundwater and integrity, which possess both clear data range descriptions and vague segmented qualitative descriptions, accurate data values ​​are often unavailable in actual engineering projects. In such cases, it's advisable to first establish partition nodes for the detection indicator based on the segmented qualitative descriptions, and then determine the risk value using inter-node interpolation. For example, regarding groundwater indicators, according to the standard "Code for Geological Investigation of Hydropower Engineering GB50287-2016," water volume <25 L / (min·10m) is classified as dry to seeping / dripping; water volume 25–125 L / (min·10m) is classified as linear flow; and water volume >125 L / (min·10m) is classified as gushing water. The specific steps are as follows:

[0065] (1) Establish the upper and lower limits of the second detection index and its correlation with the risk value. Taking the groundwater index as an example, its upper and lower limits range from dry to water inflow, and it is positively correlated with the risk value (0 to 1). Therefore, we can set the risk value to 0 when the groundwater index is dry and the risk value to 1 when the groundwater index is water inflow.

[0066] (2) Mapping and segmenting the second detection index. Based on the segmented qualitative description, the node intervals of the second detection index are directly divided. This can be done uniformly or non-uniformly, and a second mapping relationship between each interval description and the risk value is established. Taking the groundwater index as an example, the risk value can be set as 0 to 0.3 when the groundwater index is dry to seepage or dripping, 0.3 to 0.7 when the groundwater index is linear flow, and 0.7 to 1.0 when the groundwater index is gushing, thus establishing a second mapping relationship.

[0067] (3) Based on the actual geological conditions, select the detection data corresponding to the second detection index of a certain mileage and its corresponding risk value for interpolation within the interval of the second mapping relationship.

[0068] (4) Based on the results of mathematical statistics, when a sufficient amount of sample data under certain geological conditions is accumulated, this data can be used for data analysis or fitting to establish a risk value mapping that is more in line with reality. Optimization algorithms such as the least squares method can be used for data fitting.

[0069] (5) After determining the second mapping relationship, the corresponding risk value can be obtained based on the value corresponding to the detection data of the first detection index and the second mapping relationship.

[0070] In this embodiment of the application, the qualitative index threshold hierarchical mapping conversion method includes: taking the detection index with only qualitative description as the third detection index, establishing a third mapping relationship between each qualitative description of the third detection index and the risk value, and determining the corresponding risk value based on the qualitative description corresponding to the detection data of the third detection index and the third mapping relationship.

[0071] Specifically, for qualitative third-party detection indicators such as frequency variation and phase consistency of ground-penetrating radar, which typically only have image or textual interpretations and lack specific quantitative data, a given risk value can be directly proposed based on its qualitative nature. The specific steps are as follows:

[0072] (1) Mapping and segmenting the third detection indicator. Based on the qualitative description, the risk value is divided into corresponding segments. For example, if the phase consistency description of the ground-penetrating radar has three lines: obvious phase axis misalignment, slight phase axis misalignment, and basically continuous phase axis, then the risk value is divided into three segments.

[0073] (2) Based on the actual geological conditions, determine and select the specific risk values ​​for the risk value sections, and establish a third mapping relationship between the qualitative descriptions of the third detection index and the risk values. For example, the amplitude intensity description of the ground-penetrating radar is significantly enhanced, slightly enhanced, and unchanged, with risk values ​​of 0.6, 0.4, and 0.1 respectively; the frequency change description of the ground-penetrating radar is significantly changed and the waveform is chaotic, slightly changed, and unchanged, with risk values ​​of 0.6, 0.4, and 0.1 respectively; the phase consistency description of the ground-penetrating radar is significantly discontinuous in phase axis, slightly discontinuous in phase axis, and basically continuous in phase axis, with risk values ​​of 0.5, 0.3, and 0.1 respectively.

[0074] (3) Statistically adjust the risk values ​​in the above steps. The risk values ​​in the above steps are subjective. According to the results of mathematical statistics, when enough sample data under a certain geological condition is accumulated, these data can be used for data analysis to establish a more realistic mapping value for the risk values.

[0075] (4) After determining the third mapping relationship, the corresponding risk value can be obtained based on the qualitative description corresponding to the detection data of the third detection index and the third mapping relationship.

[0076] In this embodiment of the application, the detection interval is divided into multiple segments according to the coverage range of different detection data in each detection index, specifically including:

[0077] The coverage range of different detection data for each detection indicator in the tunnel is determined. There are endpoints between the coverage ranges of different detection data. The location of each endpoint of each detection indicator is used as a segmentation point. The detection interval is divided into multiple segments according to the segmentation points.

[0078] Specifically, this application embodiment divides all detection indicators corresponding to the adverse geological problems to be analyzed into segments based on different data and coverage of different detection indicators, resulting in multiple multi-source data indicator fusion calculation segments. Please refer to... Figure 2 Suppose that the detection indicators for a certain adverse geological problem include indicator A and indicator B. Indicator A has three different detection data, namely A1, A2 and A3, which cover different locations of the tunnel and have endpoints between their coverage areas. Indicator B has four different detection data, namely B1, B2, B3 and B4, which also cover different locations of the tunnel and have endpoints between their coverage areas. Based on the tunnel face, indicators A and B are aligned, and then traversed along the tunnel excavation direction. When the endpoint corresponding to any detection indicator is reached, the location of that endpoint is taken as a segmentation point. After the traversal is completed, the detection interval is divided into 6 segments based on all segmentation points.

[0079] In practical applications, due to differences in the detection ranges of different detection indicators, the detection data and their risk values ​​at different locations of the same detection indicator are also different. Therefore, this application embodiment divides the data into segments, which can ensure the consistency of the values ​​of each indicator within the same segment during data fusion, laying the foundation for multi-source data fusion.

[0080] In this embodiment of the application, the weight of each detection index is determined by using the analytic hierarchy process to calculate the first weight of each detection index, using the Huber weighting method to calculate the second weight of each detection index, and using the average of the first weight and the second weight as the weight of each detection index.

[0081] Specifically, in determining the weights of each detection indicator, the principle of basing the calculation on direct information such as geological information and drilling results, supplemented by indirect information such as geophysical exploration and experimental results, is followed. A combination of the Analytic Hierarchy Process (AHP) and the Huber weighted method is used to calculate the indicator weights. The AHP considers the subjective attribution of importance to the indicators, while the Huber weighted method focuses on the discreteness of the dataset itself. The combination of these two methods complements each other, further improving the accuracy and scientific rigor of the final indicator weights.

[0082] In practical applications, the Analytic Hierarchy Process (AHP) includes: establishing a hierarchical structure model of the detection indicators and determining the quantitative scale for judgment; performing pairwise comparisons and scoring of the detection indicators based on the quantitative scale; constructing a judgment matrix based on the scoring results; and calculating the first weight of each detection indicator based on the judgment matrix. Considering the inherent subjectivity of AHP, the weights obtained from the four commonly used AHP weight calculation methods (geometric mean, arithmetic mean, eigenvector method, and least squares method) can be averaged to reduce the result bias caused by the subjectivity of scale assignment and obtain the first weight of each detection indicator.

[0083] The Huber weighting method defines a threshold (usually estimated using standard deviation), within which the least squares method is used; beyond the threshold, a linear relationship of absolute residuals is applied. This preserves a degree of sensitivity to outliers without causing excessive susceptibility. In practical applications, a dataset containing detection indicators and adverse geological problems is constructed, and then the second weight of each detection indicator in the dataset is determined.

[0084] The weight of each detection index can be obtained by calculating the average of the first and second weights. By fusing these two weighting methods, the different sensitivities and accuracy of various forecasting methods for different types of adverse geological problems are taken into account. This solves the problems of significant differences in the physical meaning of different detection indices, varying sensitivities of different indices for the same adverse geological problem, and different sensitivities of the same indices for different adverse geological problems.

[0085] S104. Based on the risk value and weight of each detection indicator, and using the weighted average method, fuzzy fusion theory, Laplace fusion theory, Bayesian averaging method and robust statistical method, calculate the average risk index of each section, and determine whether the unfavorable geological problems to be analyzed will occur in each section based on the average risk index.

[0086] In this embodiment of the application, the method for calculating the average risk index includes:

[0087] The first risk index for each segment is calculated using the weighted average method, the second risk index for each segment is calculated using fuzzy fusion theory, the third risk index for each segment is calculated using Laplace fusion theory, the fourth risk index for each segment is calculated using the Bayesian average method, and the fifth risk index for each segment is calculated using robust statistics. The average risk index is obtained by averaging the first, second, third, fourth, and fifth risk indices.

[0088] Among them, the weighted average method, as one of the foundations of multi-source data fusion technology, integrates the data of various geological indicators by assigning appropriate weights to different data sources, including information from different data sources such as geological exploration, advanced drilling, ground-penetrating radar and seismic exploration, in order to establish a comprehensive geological condition model. By substituting the risk value and weight of the corresponding detection indicator into the geological condition model, the corresponding first risk index can be obtained.

[0089] The application of fuzzy fusion theory makes it possible to handle the uncertainty and fuzziness in geological data. In tunnel engineering, changes in geological conditions are often accompanied by uncertainty. Fuzzy fusion theory can effectively fuse and reason about this fuzzy information to more accurately predict potential geological problems encountered during tunnel construction. By using the risk values ​​and weights of corresponding detection indicators to perform fuzzy calculations, relevant information is summarized and linked together to obtain the corresponding second risk index.

[0090] Laplace fusion theory further optimizes the fusion process of multiple geological indicators by considering the correlation and weights among the indicators, thus improving the accuracy and reliability of the prediction model. The corresponding third risk index is obtained by using the risk values ​​and weights of the corresponding detection indicators, constructing a Laplace matrix, and performing a summation operation.

[0091] The Bayesian averaging method integrates prior information and dynamically adjusts the prediction model to adapt to constantly changing geological conditions. By substituting the risk values ​​and weights of the corresponding detection indicators into the prediction model, the corresponding fourth risk index can be obtained.

[0092] The application of robust statistical methods ensures the robustness and reliability of the fusion results. When faced with possible anomalies in geological data, robust statistical methods can effectively identify and handle outliers to ensure that the final geological prediction results are not affected by the outliers. By using the risk values ​​and weights of the corresponding detection indicators and comprehensively adopting the Huber weighting function and weighted average, the corresponding fifth risk index is obtained.

[0093] Finally, the average risk index of the above five risk indices is obtained by averaging them. The average risk index of each section is a value between 0 and 1. When the value is close to 0, it indicates that the probability of the unfavorable geological problem to be analyzed in the section is lower, and when the value is close to 1, it indicates that the probability of the unfavorable geological problem to be analyzed in the section is higher.

[0094] Based on the above technical solution, this embodiment also proposes a multi-source data fusion tunnel adverse geological problem analysis device, which is used to implement the steps of the multi-source data fusion tunnel adverse geological problem analysis method as described in the embodiment of this application.

[0095] It is understood that since the multi-source data fusion tunnel adverse geological problem analysis device described in this application embodiment is a device for implementing the multi-source data fusion tunnel adverse geological problem analysis method described in the embodiment, the device disclosed in the embodiment is relatively simple to describe because it corresponds to the method disclosed in the embodiment. For relevant parts, please refer to the description of the method, and it will not be repeated here.

[0096] In summary, the multi-source data fusion method and apparatus for analyzing adverse geological problems in tunnels provided in this application takes geological information as its core, constructs a database of information on the entire tunnel excavation process, and establishes a set of indicator systems for adverse geological problems in tunnel engineering. When adverse geological problems need to be analyzed, relevant detection indicators are determined from the indicator system, and detection data corresponding to each detection indicator is obtained from the database of information on the entire process. Based on the obtained detection data, risk value conversion and alignment processing are performed. Then, information technology is introduced, and mathematical geological theory is applied to carry out matrix fusion calculation of multi-source data to achieve comprehensive analysis of multi-source data fusion, obtain the analysis results of adverse geological problems in each section, realize the optimization of complementary and redundant information in space and time within the indicator system, obtain a consistent interpretation of the predicted object, and improve the accuracy of adverse geological problem analysis.

Claims

1. A method for analyzing adverse geological problems in tunnels using multi-source data fusion, characterized in that, The method includes: Establish a comprehensive information database for tunnel construction and an indicator system for adverse geological problems; Identify the current unfavorable geological problems to be analyzed in the tunnel, determine the detection indicators corresponding to the unfavorable geological problems to be analyzed in the unfavorable geological problem index system, and determine the detection data of each detection indicator within the detection interval in the whole process information database; The detection data of each detection indicator is converted into risk values. The detection interval is divided into multiple segments according to the coverage of different detection data in each detection indicator, and the weight of each detection indicator is determined. The detection interval is divided into multiple segments based on the coverage of different detection data in each detection indicator, specifically including: Determine the coverage range of different detection data for each detection indicator in the tunnel. There are endpoints between the coverage ranges of different detection data. The location of each endpoint of each detection indicator is used as a segmentation point. The detection interval is divided into multiple segments according to the segmentation points. Based on the risk value and weight corresponding to each detection indicator, and using the weighted average method, fuzzy fusion theory, Laplace fusion theory, Bayesian averaging method, and robust statistical method, the average risk index of each section is calculated. Based on the average risk index, it is determined whether the undesirable geological problems to be analyzed will occur in each section.

2. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The entire process information database includes basic tunnel design information, basic geological condition information, advanced exploration information, advanced geophysical exploration information, excavation and construction parameter information, test and inspection information, and monitoring and detection information. The basic design information includes tunnel morphology, span, height, excavation method, and surrounding rock grade; the basic geological condition information includes geomorphological features, vertical burial depth, stratigraphic code, lithology, saturation coefficient, rock mass integrity, groundwater pressure, and shear strength; the advanced exploration information includes backflow volume, backflow color, and karst cave conditions corresponding to advanced boreholes, advanced exploratory tunnels, and advanced exploratory wells; the advanced geophysical exploration information includes information corresponding to seismic wave methods, ground-penetrating radar methods, transient electromagnetic methods, and induced polarization methods; the excavation and construction parameter information includes information corresponding to drill-and-blast method construction records, TBM method construction records, and rock debris records; the test and testing information includes information corresponding to rock tests and tests, soil tests and tests, and water tests and tests; and the monitoring and detection information includes information corresponding to microseismic monitoring, deformation monitoring, and radioactivity detection.

3. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The adverse geological problems to be analyzed include high-stress rockbursts, high-stress large deformations, karst caves, karst water and mud inrushes, fault fracture zone collapses, fault fracture zone water and mud inrushes, large deformations in weak rock zones, weak rock zone collapses, high ground temperature heat hazards, toxic gas hazards, and radioactive hazards.

4. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The detection data of each indicator are converted into risk values, specifically including: The detection index with upper and lower limits and continuous intermediate values ​​is used as the first detection index. The value range of the first detection index is divided into multiple intervals, and a first mapping relationship between each interval and the risk value is established. A first mapping function is established between the first detection indicator and the risk value based on the first mapping relationship. The corresponding risk value is determined based on the value corresponding to the detection data of the first detection indicator and the first mapping function.

5. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The detection data of each indicator are converted into risk values, specifically including: The detection index with interval description but no specific value is used as the second detection index. The second mapping relationship between each interval description of the second detection index and the risk value is established. In the second mapping relationship, the interval description is added and the risk value is interpolated accordingly. The corresponding risk value is determined based on the description of the detection data of the second detection indicator and the second mapping relationship.

6. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The detection data of each indicator are converted into risk values, specifically including: The detection index with only qualitative description is used as the third detection index. A third mapping relationship is established between each qualitative description of the third detection index and the risk value. The corresponding risk value is determined based on the qualitative description corresponding to the detection data of the third detection index and the third mapping relationship.

7. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The weights of each detection indicator are determined by using the analytic hierarchy process (AHP) to calculate the first weight of each indicator, and the Huber weighting method to calculate the second weight of each indicator. The average of the first and second weights is then used as the weight of each detection indicator.

8. The method for analyzing adverse geological problems in tunnels using multi-source data fusion according to claim 1, characterized in that, The calculation method for the average risk index includes: The first risk index of each segment is calculated based on the weighted average method, the second risk index of each segment is calculated based on the fuzzy fusion theory, the third risk index of each segment is calculated based on the Laplace fusion theory, the fourth risk index of each segment is calculated based on the Bayesian average method, and the fifth risk index of each segment is calculated based on the robust statistical method. The average risk index is obtained by calculating the average of the first, second, third, fourth, and fifth risk indices.

9. A multi-source data fusion analysis device for tunnel adverse geological problems, characterized in that, The apparatus is used to implement the steps of the method for analyzing adverse geological problems in tunnels by multi-source data fusion as described in any one of claims 1 to 8.

Images