A method, system, device, and medium for conductivity tomography and formation identification

Abstract

The application discloses a conductivity tomography and stratum identification method, system, device and medium, relates to the technical field of fault image identification, and is characterized in that an electrode array is arranged in a detection pipe section, the detection pipe section is used as an information acquisition carrier, more detection information reflecting stratum characteristics after cutting is acquired, boundary electric signals in a mud conveying process are collected, and an electric conductivity distribution image of a detection section is reconstructed; characteristic parameters representing particle grading, block distribution and medium heterogeneity degree are extracted, including average conductivity, a dispersion coefficient, an abnormal area proportion, an abnormal block equivalent size, texture characteristics and time sequence fluctuation characteristics, stratum identification is carried out based on multidimensional image characteristics, a corresponding relationship between image characteristics and a working face stratum type is established, and online identification of the working face stratum is realized. The application can realize real-time and online identification of a slurry shield working face stratum, and improve tunneling safety and construction efficiency.

Description

Technical Field

[0001] This invention relates to the field of fault image recognition technology, specifically to a conductivity fault imaging and stratigraphic identification method, system, device, and medium. Background Technology

[0002] Slurry shield tunneling, due to its strong adaptability to complex geological formations, good face stability, and relatively controllable construction disturbance, has been widely applied in cross-river and cross-sea tunnels and underground engineering projects in complex urban environments. During construction, the soil or rock mass at the tunnel face is cut and broken by the cutterhead and enters the slurry circulation system, then transported to the surface treatment system via slurry discharge pipelines. Because different geological formations vary in material composition, particle size, hardness combination, and integrity, their particle size distribution, block content, and spatial distribution characteristics after entering the circulating slurry also differ significantly. Therefore, if the distribution state of the medium during slurry transportation can be monitored in real time and a correspondence between it and the geological formations at the tunnel face can be established, it is possible to achieve online identification of the geological formations at the tunnel face.

[0003] Existing methods for identifying geological formations at the tunnel face of slurry shield tunnels mainly include the following: First, predicting and segmenting the strata along the route based on preliminary geological survey data; second, making empirical judgments based on changes in construction parameters such as cutterhead torque, total thrust, propulsion speed, mud pressure, and slurry flow rate; third, manually sampling, screening, and observing and analyzing slurry, slag, or separated solid phase samples; and fourth, using ground-penetrating radar, advanced drilling, and other geophysical exploration methods for auxiliary detection. While these methods have certain application value in engineering practice, they also have significant limitations.

[0004] Among these, preliminary geological survey data, being prior information obtained before construction, is difficult to accurately reflect the real-time changes in locally abrupt geological formations and complex formations at the tunnel face during shield tunneling; discrimination methods based on construction parameters are essentially indirect empirical analyses, easily affected by the coupling of multiple factors such as equipment status, operating methods, and mud conditions, resulting in highly subjective identification results; while manual sampling analysis can reflect certain material composition characteristics, it usually suffers from problems such as sampling lag, insufficient continuity, and low automation; and advanced geological prediction methods typically require specialized equipment or procedures, making it difficult to accompany normal shield tunneling operations for extended periods and continuously.

[0005] Electrical impedance tomography (EIT) is an imaging technique that involves placing multiple electrodes at the boundary of the object under test, applying an excitation current, measuring the boundary voltage response, and then using a reconstruction algorithm to invert the cross-sectional conductivity distribution. This technique offers advantages such as non-invasiveness, fast response, and suitability for online monitoring, and has been applied in fields such as multiphase flow, slurry transport, and pipeline media distribution monitoring. For slurry transport processes, different particle compositions, concentration distributions, and aggregate states cause changes in the cross-sectional conductivity distribution; therefore, electrical impedance tomography can be used to characterize the state of the transported medium.

[0006] Existing slurry shield tunneling face stratum identification technologies still have the following shortcomings: Preliminary geological survey data is prior information and cannot promptly reflect local abrupt changes and complex strata variations at the tunnel face during excavation; discrimination methods based on construction parameters are significantly affected by multiple factors such as equipment conditions, operating methods, and slurry conditions, resulting in strong subjectivity and insufficient stability of identification results; manual sampling and slag sample analysis have lags, making continuous online identification difficult; advanced detection methods typically rely on specialized equipment and procedures, leading to high engineering implementation costs and making long-term application in continuous shield tunneling construction difficult. Furthermore, existing electrical impedance tomography (EIT) technology is mainly used for monitoring pipeline medium distribution or transport status, and has not yet developed an identification method that combines imaging features with particle size distribution, surrounding rock integrity, and tunnel face stratum type, thus failing to directly meet the needs of online stratum identification at slurry shield tunneling faces. Summary of the Invention

[0007] The technical problem this invention aims to solve is the inability to monitor the strata at the face of a slurry shield tunneling machine in real time, making it difficult to accurately reflect the real-time changes in locally abrupt and complex strata at the face during tunneling. Furthermore, the identification results are highly subjective, poorly adaptable to complex strata, and have low accuracy. The purpose is to provide a conductivity tomography imaging and strata identification method, system, equipment, and medium. By online monitoring and imaging reconstruction of the electrical distribution characteristics of the medium transported in the slurry circulation system, characteristic parameters reflecting particle size distribution, block distribution, and heterogeneity are extracted. A correspondence between imaging features and the strata type at the face is established, thereby achieving real-time, continuous, and objective identification of the strata at the face of a slurry shield tunneling machine. This provides a basis for adjusting construction parameters and risk warning under complex strata conditions.

[0008] This invention is achieved through the following technical solution:

[0009] The first aspect of this invention provides a method for conductivity tomography and stratigraphic identification, comprising the following specific steps:

[0010] Install detection pipe sections in the mud circulation system;

[0011] The detection pipe section is used to obtain mud information reflecting the characteristics of the cutting products at the working face;

[0012] Acquire the boundary voltage signal of the detection pipe section;

[0013] The acquired boundary electrical signals are preprocessed, and the preprocessed data is inverted based on a preset reconstruction algorithm to obtain the conductivity distribution image of the detection section.

[0014] The average conductivity, discrete coefficient, area ratio of abnormal regions, average equivalent size of abnormal clumps, texture features, and time-series average fluctuation features are extracted from the reconstructed image to construct a feature vector.

[0015] The feature vector is input into a preset recognition rule or training model to obtain the formation recognition result of the working face.

[0016] Furthermore, the acquisition and detection of the boundary voltage signal of the pipe segment includes:

[0017] Multiple electrode patches are arranged circumferentially on the outer wall of the detection pipe section. Each electrode is evenly distributed along the circumference of the pipe cross section and is used to contact the measured medium and apply excitation current and collect boundary voltage response.

[0018] An AC excitation signal is applied to a pair of excitation electrodes through an electrode array according to a preset excitation mode, and current is injected into the medium under test in the detection tube section.

[0019] While applying the excitation current, the boundary voltage signal caused by the change in the current field distribution is measured;

[0020] Repeatedly switch the excitation electrode pairs, traverse all preset excitation modes, and collect the complete boundary voltage signal of the detection tube segment under different excitation directions.

[0021] Furthermore, the preprocessing of the acquired boundary electrical signals includes:

[0022] A filtering window is constructed to filter the boundary voltage signal, and the filtered voltage value is obtained.

[0023] Obtain the reference voltage, normalize the filtered voltage value, and obtain the normalized voltage value.

[0024] Outlier removal is performed based on the normalized voltage values ​​to obtain the preprocessed boundary electrical signal.

[0025] Furthermore, the inversion of the preprocessed data based on the preset reconstruction algorithm includes:

[0026] Based on the boundary voltage signal, construct the boundary voltage measurement vector for each excitation mode;

[0027] The boundary voltage vector under the initial stable operating condition of the system is obtained as the baseline, and the differential voltage vector is constructed by combining it with the boundary voltage measurement vector at the current moment.

[0028] The detection section is discretized into finite element units, and a sensitivity matrix is ​​constructed to characterize the influence of the conductivity change of each unit on the boundary voltage response.

[0029] The relationship between boundary voltage change and conductivity change is established based on the sensitivity matrix and combined with the differential voltage vector, and the conductivity change vector is determined.

[0030] The change in conductivity is inverted using a preset reconstruction algorithm to obtain an estimated vector of conductivity change.

[0031] Obtain the initial conductivity distribution vector under the reference state, and combine it with the conductivity change estimation vector to obtain the reconstructed conductivity distribution vector at the current moment;

[0032] The conductivity distribution vector is mapped to a two-dimensional cross-sectional conductivity distribution image.

[0033] Furthermore, the average conductivity, dispersion coefficient, anomalous region area ratio, and average equivalent size of anomalous clumps are extracted from the reconstructed image, including:

[0034] Obtain N image units after discretizing the detection cross section;

[0035] A vector is constructed based on the conductivity values ​​of each unit to determine the average conductivity.

[0036] Determine the standard deviation and coefficient of variation of conductivity based on the average conductivity;

[0037] Determine abnormal units based on threshold criteria;

[0038] The number of units in the conductivity image that meet the anomaly criteria is obtained, and the area ratio of the abnormal region is determined by combining the N image units after the detection section is discretized.

[0039] Obtain the area of ​​the connected abnormal region and determine the equivalent size of the j-th abnormal cluster;

[0040] Based on the anomalous cluster data in the image, determine the average equivalent size of the anomalous clusters.

[0041] Furthermore, extracting texture features from the reconstructed image includes:

[0042] Acquire conductivity images, perform grayscale quantization on the images, and map the conductivity values ​​to a finite number of gray levels;

[0043] Select spatial distance and orientation parameters, count the number of times the gray value combination of pixel pairs that meet the parameter conditions occurs, and construct a gray-level co-occurrence matrix;

[0044] The gray-level co-occurrence matrix is ​​normalized to obtain the normalized gray-level co-occurrence matrix;

[0045] Calculate at least one of the following texture features based on the normalized gray-level co-occurrence matrix: contrast, energy, entropy, and homogeneity;

[0046] The texture features are output to characterize the degree of local variation, uniformity, or tissue structure features of the conductivity image.

[0047] Furthermore, extracting temporal average fluctuation features from the reconstructed image includes:

[0048] The conductivity distribution images at two adjacent time points are obtained, and the average change amplitude of the images at adjacent time points is determined to characterize the dynamic fluctuation of the formation response.

[0049] If we analyze the intensity of fluctuations over a consecutive T time intervals, we can obtain the mean of the time-series fluctuations.

[0050] A second aspect of the present invention provides a conductivity tomography and stratigraphic identification system, comprising:

[0051] The detection pipe section, located in the mud circulation system, serves as the detection area for boundary electrical signal acquisition and conductivity tomography.

[0052] An electrode array module, disposed on the detection tube segment, is used to apply an excitation current to the medium under test and acquire the boundary voltage response;

[0053] The excitation and signal acquisition module is used to apply AC excitation signals to the electrode pairs according to a preset excitation mode, and to acquire the boundary voltage signals corresponding to other measurement electrodes;

[0054] The data processing and image reconstruction module is used to preprocess the acquired boundary electrical signals, and invert the preprocessed data based on the preset reconstruction algorithm to obtain the conductivity distribution image of the detection section.

[0055] The feature extraction and recognition module is used to extract average conductivity, discrete coefficient, anomalous area ratio, average equivalent size of anomalous clumps, texture features, and time series average fluctuation features from the reconstructed image to construct a feature vector; the feature vector is input into a preset recognition rule or training model to obtain the formation recognition result of the working face;

[0056] The results output module is used to output the results of the working face strata identification, the results of the surrounding rock integrity evaluation, and the early warning information of strata change.

[0057] A third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a conductivity tomography and stratigraphic identification method.

[0058] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a conductivity tomography and stratigraphic identification method.

[0059] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0060] This invention utilizes an electrode array deployed around the outer periphery of the detection pipe section in the mud circulation system to collect boundary electrical signals during mud transport. These signals are then sequentially processed through signal filtering, normalization, outlier removal, baseline correction, image reconstruction, feature extraction, and identification analysis. This establishes a geological formation identification technology chain at the tunnel face, consisting of "boundary electrical signals—conductivity distribution image—feature parameters—stratum type," enabling online identification and continuous discrimination of the geological formation at the slurry shield tunnel face. Because this invention does not rely solely on prior geological data, construction experience, or manual sampling, but rather on imaging analysis based on the real-time electrical response of the mud transport medium, it can more promptly reflect changes in the material composition and spatial distribution after tunnel face cutting, improving the real-time nature and continuity of geological formation identification.

[0061] This invention extracts parameters such as average conductivity, dispersion coefficient, anomalous area ratio, equivalent size of anomalous clumps, texture features, and temporal fluctuation features from the conductivity distribution image of the detection section, and combines them with preset recognition rules or training models for discrimination. This allows for the quantitative expression of image feature differences corresponding to different types of strata, such as fine-grained soft soil strata, rock-fine-grained mixed strata, soft-hard composite strata, and massive rock strata. This reduces the subjectivity caused by relying solely on human experience and improves the objectivity and consistency of strata identification at the working face.

[0062] Furthermore, this invention preferably places the detection location at the slurry discharge pipeline, allowing the detected object to more directly reflect the particle size distribution, block content, and degree of heterogeneity of the medium after cutterhead cutting. Combined with image reconstruction results and connectivity feature analysis of abnormal areas, it can identify localized anomalous clumps, massive rock masses, or changes in complex strata. Therefore, this invention not only enables the identification of conventional homogeneous strata but also improves its adaptability to complex strata, abruptly changed strata, and strata with significant heterogeneity.

[0063] Meanwhile, this invention employs electrical impedance tomography to detect the slurry transport medium without interrupting the normal tunnel boring machine (TBM) excavation process. It enables non-invasive detection and continuous monitoring during slurry circulation, demonstrating strong online application capabilities. Based on the temporal analysis results of continuously reconstructed images, it can also output information on geological changes and anomalies, providing a basis for adjusting slurry shield tunneling parameters, identifying risks at the tunnel face, and controlling construction.

[0064] In summary, this invention achieves online acquisition, continuous characterization, and objective discrimination of strata information at the working face of a slurry shield tunnel through the coordinated operation of electrode array detection, boundary electrical signal preprocessing, conductivity image reconstruction, and feature extraction and recognition analysis. It has technical advantages such as good real-time performance, strong continuity, low subjectivity, and applicability to complex strata identification. Attached Figure Description

[0065] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0066] Figure 1 This is a diagram of the stratum identification system at the tunnel face of a slurry shield tunneling machine according to an embodiment of the present invention;

[0067] Figure 2 This is a diagram of the data processing and image reconstruction module in an embodiment of the present invention;

[0068] Figure 3 This is a schematic diagram of the arrangement of circumferential electrode patches on the detection tube segment in an embodiment of the present invention;

[0069] Figure 4 This is a schematic diagram of the signal acquisition technology in an embodiment of the present invention. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0071] Example 1:

[0072] like Figure 1 As shown, this embodiment provides a conductivity tomography and formation identification method, including the following specific steps: setting up a detection pipe section in the mud circulation system; acquiring mud information reflecting the characteristics of the cutting products at the working face through the detection pipe section; collecting boundary voltage signals of the detection pipe section; preprocessing the collected boundary electrical signals, and inverting the preprocessed data based on a preset reconstruction algorithm to obtain a conductivity distribution image of the detection section; extracting average conductivity, dispersion coefficient, anomalous area ratio, average equivalent size of anomalous clumps, texture features, and time-series average fluctuation features from the reconstructed image to construct a feature vector; inputting the feature vector into a preset identification rule or training model to obtain the formation identification result at the working face.

[0073] In this implementation, a detection pipe section with a responsive correlation to the formation information at the working face is selected in the mud circulation system, preferably located in the slurry discharge pipeline, so that the detection object can more directly reflect the composition and particle distribution of the formation material after cutterhead cutting. Multiple electrode patches are arranged circumferentially around the outer periphery of the detection pipe section. Each electrode is connected to the excitation and signal acquisition device and fixedly connected to the pipe section through insulation encapsulation and a protective structure. During the detection process, the excitation and acquisition device applies an AC excitation signal to the corresponding electrode according to a preset excitation method and simultaneously acquires the boundary voltage response of other measuring electrodes, thereby forming boundary electrical data characterizing the internal electrical distribution of the mud medium.

[0074] Due to differences in particle size distribution, solid-liquid composition, block content, soft-hard combination, and surrounding rock integrity among different tunnel faces during slurry shield tunneling, the electrical conductivity distribution within the slurry medium exhibits varying spatial and temporal characteristics after being cut by the cutterhead and entering the slurry circulation system. Therefore, this embodiment filters, normalizes, removes outliers, and corrects the baseline of the acquired boundary electrical signals. Furthermore, it incorporates information on the slurry's basic conductivity, temperature, flow velocity, or flow rate for condition correction to reduce the impact of external condition fluctuations on the identification results. Based on this, an electrical impedance tomography reconstruction algorithm is used to invert and reconstruct the conductivity distribution of the detected cross-section, obtaining a two-dimensional cross-sectional conductivity image at the corresponding time.

[0075] After obtaining the conductivity image, this embodiment further extracts features from the image, including the average conductivity of the cross section, the dispersion of conductivity distribution, the area ratio of abnormal regions, the number and equivalent size of abnormal clumps, image texture features, and differences between images at consecutive time points. These features characterize the content of fine particles, the distribution of coarse particles, the state of block aggregation, and the degree of heterogeneity of the medium in the mud. Subsequently, the extracted feature parameters are input into a preset recognition rule or classification model to establish a correspondence between imaging features and the strata type at the tunnel face, thereby identifying the strata at the tunnel face. The identified strata types include, but are not limited to, fine-grained soft soil strata, mixed rock and fine-grained strata, soft-hard composite strata, and massive rock strata. Further evaluation of the surrounding rock integrity can be conducted based on this. When the recognition results show that the strata at the tunnel face are transforming from relatively homogeneous strata to composite strata or massive rock strata, or when the range and dispersion of abnormal areas in the image significantly increase, the strata change trend and anomaly warning information can be output simultaneously, providing a basis for adjusting shield tunneling construction parameters and controlling risks.

[0076] Example 2:

[0077] This embodiment further illustrates a conductivity tomography and stratigraphic identification method based on Embodiment 1.

[0078] The process of acquiring boundary voltage signals of the test pipe section includes: arranging multiple electrode patches circumferentially on the outer wall of the test pipe section, with each electrode evenly distributed along the circumference of the pipe cross-section, for contacting the tested medium and applying excitation current and acquiring boundary voltage response; applying an AC excitation signal to a pair of excitation electrodes through the electrode array according to a preset excitation mode, and injecting current into the tested medium within the test pipe section; simultaneously measuring the boundary voltage signal caused by changes in the current field distribution while applying the excitation current; repeatedly switching the excitation electrode pair, traversing all preset excitation modes, and acquiring complete boundary voltage signals of the test pipe section under different excitation directions.

[0079] like Figure 2 As shown, the acquired boundary electrical signals are preprocessed, and the preprocessed data is inverted based on a preset reconstruction algorithm to obtain the conductivity distribution image of the detection cross section, specifically including:

[0080] The acquired boundary electrical signals are filtered, normalized, outlier removed, and baseline corrected. The conductivity distribution image of the detection cross-section is then obtained based on a preset reconstruction algorithm. First, the boundary voltage measurement vectors for each excitation mode are acquired.

[0081] ;

[0082] Where m is the total number of measurement data. Let T be the boundary voltage value corresponding to the i-th measurement channel, and T be the transpose symbol.

[0083] Then, moving average filtering or median filtering is used to denoise the original acquired boundary voltage signal. The moving average filtering can be expressed as:

[0084] ;

[0085] in, The voltage value after filtering at the k-th sampling point. is the original voltage value of the (k+j)th sampling point within the filtering window, n is the half-width of the sliding window, and 2n+1 is the length of the filtering window.

[0086] The filtered voltage value is then normalized.

[0087] ;

[0088] or ;

[0089] in, This is the normalized voltage value. For reference voltage value, and These are the minimum and maximum filtered voltage values ​​in the current dataset, respectively.

[0090] Outlier removal is performed based on the normalized voltage values. In this embodiment, the 3σ criterion or box plot method is used to remove outliers, where the outlier criterion is as follows:

[0091] ;

[0092] in, The mean of the normalized voltage data, This represents the standard deviation of the normalized voltage data.

[0093] Based on this, the boundary voltage vector under the initial steady-state condition of the system is used Using the baseline and combining it with the boundary voltage measurement vector at the current moment, construct the differential voltage vector:

[0094] ;

[0095] in, It is a differential voltage vector. This is the boundary voltage measurement vector at the current moment. This is the boundary voltage measurement vector under the reference operating condition.

[0096] After discretizing the detection section into finite element elements, a sensitivity matrix is ​​constructed to characterize the influence of conductivity changes in each element on the boundary voltage response. Based on the sensitivity matrix and combined with the differential voltage vector, the relationship between boundary voltage changes and conductivity changes is established, and the conductivity change vector is determined.

[0097] ;

[0098] in, This is a sensitivity matrix used to characterize the degree to which changes in the conductivity of each unit cell affect the boundary voltage response. Let be the vector of change in conductivity to be determined.

[0099] Furthermore, a preset reconstruction algorithm is used to invert the change in conductivity, resulting in an estimated vector of conductivity change. In this embodiment, the Tikhonov regularization method is preferably used.

[0100] ;

[0101] in, This is the estimated vector of conductivity change obtained from the inversion. Let λ be the transpose of the sensitivity matrix S, and λ be the regularization parameter. It is an identity matrix.

[0102] Obtain the initial conductivity distribution vector under the reference state, and combine it with the conductivity change estimation vector to obtain the reconstructed conductivity distribution vector at the current time:

[0103] ;

[0104] in, This is the conductivity distribution vector reconstructed at the current moment. This is the initial conductivity distribution vector under the reference state.

[0105] The conductivity distribution vector is then mapped into a two-dimensional cross-sectional conductivity distribution image for subsequent analysis of particle size distribution characteristics, block distribution characteristics, and formation identification at the working face.

[0106] Parameters such as average conductivity, coefficient of variation, area ratio of anomalous regions, equivalent size of anomalous clumps, texture features, and time-series fluctuation features are extracted from the reconstructed image, and the stratigraphic type of the tunnel face is determined according to preset recognition rules or trained models. In other words, feature extraction and stratigraphic identification are performed on the reconstructed conductivity distribution image.

[0107] First, extract the average conductivity, dispersion coefficient, anomalous region area ratio, and average equivalent size of anomalous clumps from the reconstructed image, including:

[0108] Suppose that the discretized cross section has N image units, and the conductivity values ​​of each unit form a vector:

[0109] ;

[0110] Based on the conductivity values ​​of each unit, a vector is constructed to determine the average conductivity. The average conductivity is then:

[0111] ;

[0112] in, Average conductivity is used to characterize the overall conductivity level of the test cross section.

[0113] Based on the average conductivity, the standard deviation of conductivity is determined as follows:

[0114] ;

[0115] Based on the standard deviation of conductivity, the coefficient of variation is determined as follows:

[0116] ;

[0117] in, The standard deviation of conductivity, is the discrete coefficient, used to characterize the degree of non-uniformity in conductivity distribution. The larger the value, the more pronounced the heterogeneity within the detection cross-section.

[0118] Let the number of cells in the conductivity image that satisfy the anomaly criteria be . The total number of image units is The area ratio of the abnormal region is:

[0119] ;

[0120] Abnormal units can be determined using threshold criteria:

[0121] ;

[0122] Where η is the preset threshold coefficient.

[0123] For a region with abnormal connectivity, obtain its area, and let its area be... Then the equivalent size of the j-th anomalous cluster can be expressed as:

[0124] ;

[0125] in, Let be the equivalent diameter of the j-th anomalous cluster. This represents the area corresponding to the anomalous cluster.

[0126] If there are M anomalous blobs in the image, their average equivalent size can be expressed as:

[0127] ;

[0128] in, This represents the average equivalent size of the abnormal clusters.

[0129] Secondly, texture features are extracted from the reconstructed image, including: acquiring the conductivity image, performing grayscale quantization on the image, and mapping the conductivity values ​​to a finite number of gray levels; selecting spatial distance and orientation parameters, counting the occurrence frequency of grayscale value combinations of pixel pairs that satisfy the parameter conditions, and constructing a gray-level co-occurrence matrix; normalizing the gray-level co-occurrence matrix to obtain a normalized gray-level co-occurrence matrix; and calculating at least one of the following texture features based on the normalized gray-level co-occurrence matrix, including: contrast, energy, entropy, and homogeneity; let the normalized gray-level co-occurrence matrix be p(i,j), then the calculation process is as follows:

[0130] Contrast: ;

[0131] energy: ;

[0132] entropy: ;

[0133] Homogeneity: ;

[0134] The above texture features are output to characterize the degree of local variation, uniformity, and tissue structure features of the conductivity image.

[0135] Finally, temporal average fluctuation features are extracted from the reconstructed images, including:

[0136] Obtain the conductivity distribution images at two adjacent time points, let the conductivity distribution images at two adjacent time points be respectively... and Then the time series difference can be expressed as:

[0137] ;

[0138] Wherein, Δt is the average change amplitude of images at adjacent time points, used to characterize the degree of dynamic fluctuation in the formation response.

[0139] If we analyze the intensity of fluctuations over a consecutive T time periods, we can define the mean of the time series fluctuations as:

[0140] ;

[0141] in, It represents the time-series average fluctuation characteristics.

[0142] After obtaining the above characteristic parameters, namely, the average conductivity, the coefficient of variation, the area ratio of the anomalous region, the average equivalent size of the anomalous clusters, the texture features (contrast, energy, entropy, and homogeneity), and the time-series average fluctuation features, a feature vector is constructed:

[0143] ;

[0144] Where F is the feature vector used for stratigraphic identification.

[0145] Input the feature vector into the preset recognition rules or training model to obtain the formation recognition results at the tunnel face:

[0146] ;

[0147] in, Y is the identification function, and Y is the identification output. Y can correspond to one of the following: fine-grained soft soil strata, rock-fine-grained mixed strata, soft-hard composite strata, or massive rock strata, thereby realizing the identification of the strata type at the tunnel face.

[0148] Example 3:

[0149] This embodiment provides a conductivity tomography and strata identification system. Through a modular architecture design, this device implements any one of the conductivity tomography and strata identification methods described in Embodiments 1 and 2 above. The system includes a detection pipe section, an electrode array, an excitation and signal acquisition device, a data processing and image reconstruction device, a feature extraction and recognition device, and a result output device. These devices are connected sequentially and work collaboratively to complete the entire process of mud boundary electrical signal acquisition, conductivity image reconstruction, image feature analysis, and tunnel face strata identification. Through the above technical solution, this invention combines the electrical imaging results during mud transport with tunnel face strata identification, establishing an identification chain of "boundary electrical signal—cross-sectional conductivity distribution—image feature parameters—tunnel face strata type," thereby achieving online, continuous, and objective identification of the strata at the tunnel face of a slurry shield tunnel. It includes:

[0150] The detection pipe section is set in the mud circulation system and is used as the detection area for boundary electrical signal acquisition and conductivity tomography. The detection pipe section can be set in the slurry discharge pipe, slurry inlet pipe or main mud circulation channel, preferably in the slurry discharge pipe, so as to obtain mud information that better reflects the characteristics of the cutting products at the working face.

[0151] An electrode array module, mounted on the detection pipe section, is used to apply an excitation current to the measured medium and acquire the boundary voltage response. That is, multiple electrode patches are arranged circumferentially on the outer wall of the detection pipe section, with each electrode evenly distributed along the circumference of the pipe cross-section. The number of electrodes can be set to 8, 12, 16, or more depending on the required detection accuracy, preferably 16. Figure 3 As shown.

[0152] The excitation and signal acquisition module is used to apply AC excitation signals to the electrode pairs according to a preset excitation mode and to acquire the boundary voltage signals corresponding to other measuring electrodes. Specifically, it is electrically connected to the electrode array and is used to apply AC excitation signals to the electrode pairs according to a preset excitation mode and to acquire the boundary voltage signals corresponding to other measuring electrodes. The excitation method can employ one or more combinations of adjacent excitation, opposed excitation, or interval excitation. The system injects a small, safe excitation current into the internal material through a pair of electrodes installed at the boundary of an object (e.g., the outer wall of a pipe or reactor). The injected current flows within the internal material, forming a current field. The current path (e.g., ...) Figure 4 The conductivity distribution (as shown by the curved lines in the middle) changes depending on the different materials or states within the system. The system then measures the resulting boundary voltage between all the remaining unused electrode pairs.

[0153] The data processing and image reconstruction module is used to preprocess the acquired boundary electrical signals, and invert the preprocessed data based on the preset reconstruction algorithm to obtain the conductivity distribution image of the detection section.

[0154] The feature extraction and recognition module is used to extract average conductivity, discrete coefficient, anomalous area ratio, average equivalent size of anomalous clumps, texture features, and time series average fluctuation features from the reconstructed image to construct a feature vector; the feature vector is then input into a preset recognition rule or training model to obtain the formation recognition result at the working face;

[0155] The results output module is used to output the results of the working face strata identification, the results of the surrounding rock integrity evaluation, and the early warning information of strata change.

[0156] As one possible implementation, this embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a conductivity tomography and stratigraphic identification method.

[0157] As one possible implementation, this embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a conductivity tomography and stratigraphic identification method.

[0158] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for conductivity tomography and stratigraphic identification, characterized in that, The specific steps include the following: Install detection pipe sections in the mud circulation system; The detection pipe section is used to obtain mud information reflecting the characteristics of the cutting products at the working face; Acquire the boundary voltage signal of the detection pipe section; The acquired boundary electrical signals are preprocessed, and the preprocessed data is inverted based on a preset reconstruction algorithm to obtain the conductivity distribution image of the detection section. The average conductivity, discrete coefficient, area ratio of abnormal regions, average equivalent size of abnormal clumps, texture features, and time-series average fluctuation features are extracted from the reconstructed image to construct a feature vector. The feature vector is input into a preset recognition rule or training model to obtain the formation recognition result of the working face.

2. The conductivity tomography and stratigraphic identification method according to claim 1, characterized in that, The acquisition and detection of the boundary voltage signal of the pipe section includes: Multiple electrode patches are arranged circumferentially on the outer wall of the detection pipe section. Each electrode is evenly distributed along the circumference of the pipe cross section and is used to contact the measured medium and apply excitation current and collect boundary voltage response. An AC excitation signal is applied to a pair of excitation electrodes through an electrode array according to a preset excitation mode, and current is injected into the medium under test in the detection tube section. While applying the excitation current, the boundary voltage signal caused by the change in the current field distribution is measured; Repeatedly switch the excitation electrode pairs, traverse all preset excitation modes, and collect the complete boundary voltage signal of the detection tube segment under different excitation directions.

3. The conductivity tomography and stratigraphic identification method according to claim 1, characterized in that, The preprocessing of the acquired boundary electrical signals includes: A filtering window is constructed to filter the boundary voltage signal, and the filtered voltage value is obtained. Obtain the reference voltage, normalize the filtered voltage value, and obtain the normalized voltage value. Outlier removal is performed based on the normalized voltage values ​​to obtain the preprocessed boundary electrical signal.

4. The conductivity tomography and stratigraphic identification method according to claim 1, characterized in that, The inversion of preprocessed data based on a preset reconstruction algorithm includes: Based on the boundary voltage signal, construct the boundary voltage measurement vector for each excitation mode; The boundary voltage vector under the initial stable operating condition of the system is obtained as the baseline, and the differential voltage vector is constructed by combining it with the boundary voltage measurement vector at the current moment. The detection section is discretized into finite element units, and a sensitivity matrix is ​​constructed to characterize the influence of the conductivity change of each unit on the boundary voltage response. The relationship between boundary voltage change and conductivity change is established based on the sensitivity matrix and combined with the differential voltage vector, and the conductivity change vector is determined. The change in conductivity is inverted using a preset reconstruction algorithm to obtain an estimated vector of conductivity change. Obtain the initial conductivity distribution vector under the reference state, and combine it with the conductivity change estimation vector to obtain the reconstructed conductivity distribution vector at the current moment; The conductivity distribution vector is mapped to a two-dimensional cross-sectional conductivity distribution image.

5. The conductivity tomography and stratigraphic identification method according to claim 1, characterized in that, The average conductivity, dispersion coefficient, anomalous region area ratio, and average equivalent size of anomalous clumps are extracted from the reconstructed image, including: Obtain N image units after discretizing the detection cross section; A vector is constructed based on the conductivity values ​​of each unit to determine the average conductivity. Determine the standard deviation and coefficient of variation of conductivity based on the average conductivity; Determine abnormal units based on threshold criteria; The number of units in the conductivity image that meet the anomaly criteria is obtained, and the area ratio of the abnormal region is determined by combining the N image units after the detection section is discretized. Obtain the area of ​​the connected abnormal region and determine the equivalent size of the j-th abnormal cluster; Based on the anomalous cluster data in the image, determine the average equivalent size of the anomalous clusters.

6. The conductivity tomography and stratigraphic identification method according to claim 5, characterized in that, Extracting texture features from the reconstructed image includes: Acquire conductivity images, perform grayscale quantization on the images, and map the conductivity values ​​to a finite number of gray levels; Select spatial distance and orientation parameters, count the number of times the gray value combination of pixel pairs that meet the parameter conditions occurs, and construct a gray-level co-occurrence matrix; The gray-level co-occurrence matrix is ​​normalized to obtain the normalized gray-level co-occurrence matrix; Calculate at least one of the following texture features based on the normalized gray-level co-occurrence matrix: contrast, energy, entropy, and homogeneity; The texture features are output to characterize the degree of local variation, uniformity, or tissue structure features of the conductivity image.

7. The conductivity tomography and stratigraphic identification method according to claim 6, characterized in that, Extracting temporal average fluctuation features from the reconstructed image includes: The conductivity distribution images at two adjacent time points are obtained, and the average change amplitude of the images at adjacent time points is determined to characterize the dynamic fluctuation of the formation response. If we analyze the intensity of fluctuations over a consecutive T time intervals, we can obtain the mean of the time-series fluctuations.

8. A conductivity tomography and stratigraphic identification system, characterized in that, include: The detection pipe section, located in the mud circulation system, serves as the detection area for boundary electrical signal acquisition and conductivity tomography. An electrode array module, disposed on the detection tube segment, is used to apply an excitation current to the medium under test and acquire the boundary voltage response; The excitation and signal acquisition module is used to apply AC excitation signals to the electrode pairs according to a preset excitation mode, and to acquire the boundary voltage signals corresponding to other measurement electrodes; The data processing and image reconstruction module is used to preprocess the acquired boundary electrical signals, and invert the preprocessed data based on the preset reconstruction algorithm to obtain the conductivity distribution image of the detection section. The feature extraction and recognition module is used to extract average conductivity, discrete coefficient, anomalous area ratio, average equivalent size of anomalous clumps, texture features, and time series average + fluctuation features from the reconstructed image to construct a feature vector; the feature vector is input into a preset recognition rule or training model to obtain the formation recognition result of the working face; The results output module is used to output the results of the working face strata identification, the results of the surrounding rock integrity evaluation, and the early warning information of strata change.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the conductivity tomography and stratigraphic identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the conductivity tomography and stratigraphic identification method as described in any one of claims 1 to 7.

Images