Multi-dimensional resistivity water-abundance grading method for deep tunnel

Abstract

The invention relates to the technical field of geophysics, in particular to a multi-dimensional resistivity water-abundance grading method for a deep tunnel. According to the method, the water yield grade of the to-be-graded deep-buried tunnel is comprehensively predicted by adopting a multi-dimensional method in the tunnel investigation stage and the construction stage, and the water yield grade prediction is more accurate. The water yield grading method provided by the invention can effectively describe the importance degree between the parameters, and quantitatively judge the water yield grade in front of the tunnel face, so that the prediction result is more in line with the actual situation, and the construction risk caused by the uncertainty of the water yield strength is reduced.

Description

Technical Field

[0001] The invention relates to the field of geophysical technology, and in particular to a multi-dimensional resistivity water-richness classification method for deep-buried tunnels. Background Art

[0002] In recent years, with the development of high-speed railways, tunnel construction has become more and more complex, and the depth of the tunnels has become deeper and deeper, such as the Chongqing-Kunming Railway, Dali-Ruili Railway, Sichuan-Tibet Railway, which are currently under construction, and the Xinjiang-Tibet Railway and Yunnan-Tibet Railway, which are about to be built. Long and deep tunnels (over 30 meters) with complex structures often have strong water-rich sections, and mud and water surges are very likely to occur during construction and excavation, causing unnecessary property losses and casualties. The uncertainty of the water-richness level will bring great construction risks.

[0003] In the survey stage of long and deep railway tunnels, geophysical exploration methods such as airborne electromagnetic method and magnetotelluric method are often used to make a preliminary qualitative determination of water-richness. Comprehensive logging and drilling methods can quantitatively determine the strength of water-richness, but this method cannot represent the whole with a single point. In addition, in the construction stage, transient electromagnetic method is often used to qualitatively determine water-richness. The above methods can only preliminarily determine whether the construction section contains water, but cannot accurately determine the strength of water-richness, and thus cannot accurately predict tunnel geological disasters such as sudden mud and water inrush. Therefore, in response to the above problems, this paper combines airborne geophysical exploration, magnetotelluric, comprehensive logging methods, tunnel transient electromagnetic, borehole transient electromagnetic and other methods in the tunnel survey stage to quantitatively determine the water-richness level in front of the tunnel face, reducing the construction risks caused by the uncertainty of the strength of water-richness.

[0004] The Chinese patent with publication number CN 115598713 A and publication date of January 13, 2023 discloses a method for predicting the water-richness of sandstone layers outside a borehole based on a fully convolutional neural network. The invention discloses a method for predicting the water-richness of sandstone layers outside a borehole based on a fully convolutional neural network, comprising: obtaining measured original data; the measured original data is the induced electromotive force of the response of the water-bearing sandstone geological body outside the borehole collected by the receiving point in the borehole according to the sampling time series; extracting features from the measured original data to generate feature information data; inputting the feature information data into the fully convolutional neural network, and outputting the water-richness classification prediction information of the water-bearing sandstone geological body outside the borehole according to the preset water-richness level classification standard; the water-richness classification prediction information is a water-richness category, including: water-free surrounding rock, weak water-rich surrounding rock, medium water-rich surrounding rock, strong water-rich surrounding rock and extremely strong water-rich surrounding rock. It can effectively improve the classification prediction accuracy of the water-richness of the sandstone aquifer in the radial direction outside the borehole wall. However, this patent uses the transient electromagnetic method during the construction phase to only classify the water-richness of sandstone, and is unable to classify the water-richness of common surrounding rocks, and the classification method is single.

[0005] The Chinese patent with publication number CN 112213792 A and publication date January 12, 2021 discloses a method for predicting the water-richness of a water-bearing geological body based on transient electromagnetic method. The present invention provides a method for predicting the water-richness of a water-bearing geological body based on transient electromagnetic method, which relates to the technical field of mine safety prevention and control. The method for predicting the water-richness of a water-bearing geological body based on transient electromagnetic method comprises the following steps: step one, mine transient electromagnetic detection; step two, obtaining dual-field information data of a water-bearing geological body and water-richness information data of a borehole; step three, constructing an intelligent discrimination model for the water-richness of a water-bearing geological body; step four, predicting the water-richness of a water-bearing geological body. The present invention is applied to the detection construction and data processing and interpretation of transient electromagnetic method in mines, and uses the "double field" information of transient electromagnetic diffusion field and pseudo-seismic wave field to extract water-rich characteristics and construct sample sets, overcoming the disadvantage of single information characteristics of transient electromagnetic raw data, and using deep neural network algorithm to optimize and approximate the complex nonlinear relationship between sample data and water-rich information of water-bearing geological bodies, so as to achieve deep excavation and graded prediction of water-rich information of water-bearing geological bodies. However, the patent uses transient electromagnetic method to grade and predict the water-richness of surrounding rocks during the construction stage, and does not use a comprehensive method for grading, so the grading effect is poor.

[0006] Therefore, a water-richness classification method with stronger applicability and more accurate classification is needed nowadays. Summary of the invention

[0007] The purpose of the present invention is to overcome the problems of narrow applicability and poor grading effect in the prior art, and to provide a multi-dimensional resistivity water-richness grading method for deep buried tunnels.

[0008] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0009] A multi-dimensional resistivity water-richness classification method for a deep buried tunnel comprises the following steps:

[0010] S1: Acquire electromagnetic method data and ground comprehensive logging data of the deep-buried tunnel to be graded during the survey phase and electromagnetic data of the deep-buried tunnel to be graded during the construction phase;

[0011] S2: Inverting the electromagnetic data of the survey phase and the electromagnetic data of the construction phase by a multi-parameter joint inversion algorithm, and outputting resistivity inversion results;

[0012] S3: According to the ground comprehensive logging data and the resistivity inversion result, the water-richness classification of the deep-buried tunnel to be classified is performed, and the water-richness classification result of the deep-buried tunnel to be classified is output.

[0013] As a preferred solution of the present invention, the electromagnetic data is acquired by airborne electromagnetic method and / or magnetotelluric method and / or wide-area electromagnetic method.

[0014] As a preferred embodiment of the present invention, the electromagnetic data in the construction stage includes excavation face transient electromagnetic data or drilling transient electromagnetic data.

[0015] As a preferred solution of the present invention, the expression of the multi-parameter joint inversion algorithm in S2 is:

[0016]

[0017] in, is the resistivity inversion result, obs represents the observed data, represents the ith electromagnetic method data of the survey stage, represents the ith electromagnetic data of the tunnel during the construction phase, Represents the inversion gradient formula corresponding to the electromagnetic data in the survey stage, represents the inversion gradient formula corresponding to the electromagnetic data, * represents the complex conjugate, λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn is the regularization weight factor, λ is the regularization factor, and λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn Satisfy: λ k1 +λ k2 +···+λ km =1,λ s1 +λ s2 +···+λ sn =1, F represents the forward operator, F(m) ki represents the forward modeling data corresponding to the ith electromagnetic method data in the survey stage, F(m) si represents the forward modeling data corresponding to the i-th electromagnetic data in the construction stage, and represents the data and model covariance matrix, m and m0 are the model parameters and prior model of resistivity, respectively.

[0018] As a preferred solution of the present invention, the S3 analyzes the ground comprehensive logging data and the resistivity inversion results through a pre-constructed parallel fully connected Block module neural network, outputs a water-richness index, and classifies the water-richness according to the water-richness index.

[0019] As a preferred solution of the present invention, the parallel fully connected Block module neural network includes two parallel fully connected Block modules, a pooling layer and a fully connected layer;

[0020] The fully connected Block module includes a plurality of sequentially connected convolutional layers and fully connected Block layers, and a convolutional layer.

[0021] As a preferred embodiment of the present invention, the grading standard of the water-richness index in S3 is:

[0022] Very weak water-rich area Weak water-rich area Moderately water-rich area Strong water-rich area Ⅰ Ⅱ Ⅲ Ⅳ 0-0.25 0.25-0.5 0.5-0.75 0.75-1 .

[0023] A multi-dimensional resistivity water-richness grading device for a deep-buried tunnel comprises at least one processor and a memory in communication with the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the methods described above.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The present invention adopts a multi-dimensional method to comprehensively predict the water-richness level of the deep-buried tunnel to be graded during the tunnel survey and construction stages, and the prediction of the water-richness level is more accurate. The water-richness grading method proposed in the present invention can effectively characterize the importance of each parameter, and quantitatively determine the water-richness level in front of the tunnel face, so that the prediction results are more in line with the actual situation and reduce the construction risks caused by the uncertainty of the water-richness strength. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a schematic flow chart of a multi-dimensional resistivity water-richness classification method for a deep-buried tunnel according to Example 1 of the present invention;

[0027] Figure 2 This is a schematic flow chart of a multi-dimensional resistivity water-richness classification method for a deep-buried tunnel according to Embodiment 2 of the present invention;

[0028] Figure 3 The electromagnetic method and comprehensive logging data of the tunnel survey stage in the multi-dimensional resistivity water-richness classification method for deep-buried tunnels described in Example 2 of the present invention;

[0029] Figure 4 Transient electromagnetic data of the tunnel construction stage in a multi-dimensional resistivity water-richness classification method for a deep buried tunnel as described in Example 2 of the present invention;

[0030] Figure 5This is a schematic diagram of the structure of a parallel fully connected Block module neural network in a multi-dimensional resistivity water-richness classification method for a deep buried tunnel according to Example 2 of the present invention;

[0031] Figure 6 This is a quantitative determination result diagram of the water-richness grade in a multi-dimensional resistivity water-richness classification method for a deep-buried tunnel according to Example 2 of the present invention;

[0032] Figure 7 This is a schematic structural diagram of a multi-dimensional resistivity water-richness grading device for deep buried tunnels described in Example 3 of the present invention, which utilizes the multi-dimensional resistivity water-richness grading method for deep buried tunnels described in Example 1. DETAILED DESCRIPTION

[0033] The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

[0034] Example 1

[0035] like Figure 1 As shown, a multi-dimensional resistivity water-richness classification method for deep buried tunnels includes the following steps:

[0036] S1: Acquire electromagnetic data and ground comprehensive logging data of the deep-buried tunnel to be graded during the survey phase and electromagnetic data of the deep-buried tunnel to be graded during the construction phase.

[0037] S2: Inverting the electromagnetic data of the survey phase and the electromagnetic data of the construction phase through a multi-parameter joint inversion algorithm, and outputting resistivity inversion results.

[0038] S3: According to the ground comprehensive logging data and the resistivity inversion result, the water-richness classification of the deep-buried tunnel to be classified is performed, and the water-richness classification result of the deep-buried tunnel to be classified is output.

[0039] Example 2

[0040] This embodiment is a specific implementation method of the multi-dimensional resistivity water-richness classification method for deep buried tunnels described in Example 1. Figure 2 As shown, the following steps are included:

[0041] S1: Acquire electromagnetic data and ground comprehensive logging data of the deep-buried tunnel to be graded during the survey phase and electromagnetic data of the deep-buried tunnel to be graded during the construction phase.

[0042] like Figure 3As shown, the electromagnetic data are obtained by airborne electromagnetic method and / or magnetotelluric method and / or wide-area electromagnetic method, and are recorded as resistivity 1, resistivity 2, ... resistivity n after inversion calculation.

[0043] The ground comprehensive logging data are recorded as logging data 1, logging data 2, logging data 3, ... logging data n.

[0044] The electromagnetic data in the construction phase includes transient electromagnetic data of the excavation face or transient electromagnetic data of the drilling hole; the transient electromagnetic data is as follows: Figure 4 shown.

[0045] S2: Inverting the electromagnetic data of the survey phase and the electromagnetic data of the construction phase through a multi-parameter joint inversion algorithm, and outputting resistivity inversion results.

[0046] This paper involves multi-parameter electromagnetic data in the survey stage and the construction stage. In order to achieve more accurate results, an improved multi-parameter joint inversion gradient formula is adopted. The expression of the multi-parameter joint inversion algorithm is:

[0047]

[0048] in, is the resistivity inversion result, obs represents the observed data, represents the ith electromagnetic method data of the survey stage, represents the ith electromagnetic data of the tunnel during the construction phase, Represents the inversion gradient formula corresponding to the electromagnetic data in the survey stage, represents the inversion gradient formula corresponding to the electromagnetic data, * represents the complex conjugate, λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn is the regularization weight factor, λ is the regularization factor, and λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn Satisfy: λ k1 +λ k2 +···+λ km =1,λ s1 +λ s2 +···+λ sn =1, F represents the forward operator, F(m) ki represents the forward modeling data corresponding to the ith electromagnetic method data in the survey stage, F(m) sirepresents the forward modeling data corresponding to the i-th electromagnetic data in the construction stage, and represents the data and model covariance matrix, m and m0 are the model parameters and prior model of resistivity, respectively.

[0049] S3: According to the ground comprehensive logging data and the resistivity inversion result, the water-richness classification of the deep-buried tunnel to be classified is performed, and the water-richness classification result of the deep-buried tunnel to be classified is output.

[0050] Furthermore, this embodiment proposes a neural network structure with a parallel fully connected Block module based on the difference in data structure. It is not only related in the horizontal direction but also in the vertical direction on the spatial matrix. Therefore, the parallel network structure is designed to pay attention to the characteristics of both the horizontal and vertical directions. In addition, in the fully connected Block module, each layer is directly connected to all other layers, which makes information transmission more direct and efficient. In contrast, traditional convolutional networks only have connections between adjacent layers. Specifically, this step analyzes the ground comprehensive logging data and the resistivity inversion results through a pre-constructed parallel fully connected Block module neural network, outputs a water-richness index, and classifies the water-richness according to the corresponding water-richness index.

[0051] like Figure 5 As shown, the parallel fully connected Block module neural network includes two parallel fully connected Block modules, a pooling layer and a fully connected layer;

[0052] The fully connected Block module includes a plurality of sequentially connected convolutional layers and fully connected Block layers, and a convolutional layer.

[0053] This module has several advantages: it can solve the problem of inaccurate results caused by differences in data structures and make the calculation results more accurate; in addition, it can solve the gradient disappearance problem, enhance feature propagation, encourage feature reuse, and require fewer parameters.

[0054] The water-richness of the deep-buried tunnel to be classified is classified according to the preset standard, and the corresponding water-richness classification result is output. The water-richness classification result is as follows: Figure 6 shown.

[0055] Furthermore, the grading standard of water richness index is:

[0056] Very weak water-rich area Weak water-rich area Moderately water-rich area Strong water-rich area Ⅰ Ⅱ Ⅲ Ⅳ 0-0.25 0.25-0.5 0.5-0.75 0.75-1 .

[0057] Example 3

[0058] like Figure 7As shown, a multi-dimensional resistivity water-richness classification device for deep buried tunnels includes at least one processor, a memory connected to the at least one processor, and at least one input-output interface connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the multi-dimensional resistivity water-richness classification method for deep buried tunnels described in the above embodiment. The input-output interface may include a display, a keyboard, a mouse, and a USB interface for inputting and outputting data.

[0059] Those skilled in the art can understand that: all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM), magnetic disks or optical disks, etc. Various media that can store program codes.

[0060] When the above-mentioned integrated unit of the present invention is implemented in the form of a software functional unit and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-dimensional resistivity water-richness classification method for deep buried tunnels, characterized in that: The following steps are involved: S1: Acquire electromagnetic method data and ground comprehensive logging data of the deep-buried tunnel to be graded during the survey phase and electromagnetic data of the deep-buried tunnel to be graded during the construction phase; S2: Inverting the electromagnetic data of the survey phase and the electromagnetic data of the construction phase by a multi-parameter joint inversion algorithm, and outputting resistivity inversion results; S3: According to the ground comprehensive logging data and the resistivity inversion result, the water-richness classification of the deep-buried tunnel to be classified is performed, and the water-richness classification result of the deep-buried tunnel to be classified is output.

2. A multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 1, characterized in that: The electromagnetic data are acquired by airborne electromagnetic method, magnetotelluric method and / or wide-area electromagnetic method.

3. A multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 1, characterized in that: The electromagnetic data in the construction phase includes excavation face transient electromagnetic data or drilling transient electromagnetic data.

4. A multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 1, characterized in that: The expression of the multi-parameter joint inversion algorithm in S2 is: in, is the resistivity inversion result, obs represents the observed data, represents the ith electromagnetic method data of the survey stage, represents the ith electromagnetic data of the tunnel during the construction phase, Represents the inversion gradient formula corresponding to the electromagnetic data in the survey stage, represents the inversion gradient formula corresponding to the electromagnetic data, * represents the complex conjugate, λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn is the regularization weight factor, λ is the regularization factor, and λ k1 ,λ k2 ,···,λ km and s1 ,λ s2 ,···,λ sn Satisfy: λ k1 +λ k2 +···+λ km =1,λ s1 +λ s2 +···+λ sn =1, F represents the forward operator, F(m) ki represents the forward modeling data corresponding to the ith electromagnetic method data in the survey stage, F(m) si represents the forward modeling data corresponding to the i-th electromagnetic data in the construction stage, and represents the data and model covariance matrix, m and m0 are the model parameters and prior model of resistivity, respectively.

5. The multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 1 is characterized in that: The S3 analyzes the ground comprehensive logging data and the resistivity inversion results through a pre-constructed parallel fully connected Block module neural network, outputs a water-richness index, and classifies the water-richness according to the water-richness index.

6. A multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 5, characterized in that: The parallel fully connected Block module neural network includes two parallel fully connected Block modules, a pooling layer and a fully connected layer; The fully connected Block module includes a plurality of sequentially connected convolutional layers and fully connected Block layers, and a convolutional layer.

7. A multi-dimensional resistivity water-richness classification method for deep buried tunnels according to claim 6, characterized in that: The grading standard of the water-richness index in S3 is: 。 8. A multi-dimensional resistivity water-richness classification device for deep buried tunnels, characterized in that: It includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in any one of claims 1 to 7.

Images