Method and system for identifying geological characteristics of shield tunneling based on fuzzy classification algorithm
Through fuzzy classification algorithm and principal component analysis, combined with shield machine sensor data and geological profile diagram, the geological characteristics of shield excavation are identified in real time, which solves the problem of inaccurate identification in the existing technology, improves construction efficiency and safety, and realizes the High accuracy shield excavation.
Patent Information
- Application Number
- CN202210098529.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-27
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-01-27
AI Technical Summary
The prior art cannot accurately identify the geological characteristics during shield excavation, especially in uneven soft and hard strata, which affects the safety and efficiency of shield excavation.
Using a method based on fuzzy classification algorithm, combining principal component analysis and fuzzy division coefficients and contour coefficients, we can identify geological characteristics during shield excavation in real time, and determine the geological characteristic categories and provide accurate Membership degree.
It realizes accurate identification of geological characteristics during shield excavation, improves construction efficiency and safety, ensures the accuracy of shield tunnel construction, and the recognition accuracy rate reaches more than 90%.
Smart Images

Figure CN114492038B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of shield tunneling, and more specifically, to a shield tunneling geological feature recognition method and system based on a fuzzy classification algorithm. Background Art
[0002] With the continuous acceleration of urbanization, exchanges between cities are becoming more and more frequent. Intercity rail transit systems have developed rapidly. Shield construction is widely used in the construction of urban underground tunnels due to its safety, speed and environmental protection. The shield method uses a shield cutter to cut rock and soil, and then transfers the soil to a muck truck through soil bins, screw machines, conveyor belts and other equipment, and then transports it to the outside of the tunnel by the muck truck; the excavated tunnel is assembled by a segment assembly machine and grouting is performed to form a circular shield tunnel. During the excavation of the shield machine, geological characteristics are the most important factor affecting the speed of shield excavation. Before tunnel construction, geological surveys are usually conducted to draw geological profiles of the construction area. However, the geological profile drawn by connecting the geological characteristics of the survey holes in a straight line cannot accurately reveal the geological conditions of each ring of shield excavation. In order to ensure the safe and efficient excavation of the shield, the identification of geological characteristics during the shield excavation process is crucial.
[0003] The proportion of rock strata in the shield tunneling strata has a significant impact on shield tool wear, tunneling efficiency and shield parameter settings. Setting shield parameters and shield tunneling schedules according to different rock strata proportions can ensure safe and efficient operation of the shield. In "Research on Key Technologies for Long-Distance Hard Rock Tunneling Construction under Complex Conditions" published in "Value Engineering" in 2019, Tian Chao emphasized the importance of parameter optimization for complex strata during shield tunneling. At present, the geological characteristics of the construction area are inferred from geological survey holes, and the geological characteristics of the strata where the shield tunneling is located cannot be given, and accurate geological information cannot be provided for shield tunneling in uneven soft and hard strata. Therefore, in order to obtain the geological conditions in front of the shield tunneling cutterhead and guide the tunneling arrangement and tool replacement of the shield machine, it is necessary to propose a shield tunneling geological feature recognition method based on fuzzy classification algorithm.
[0004] After searching the literature on the prior art, it was found that the prior art discloses a method and system for detecting and identifying geological features based on deep learning. This scheme uses the method of deep learning to classify geological feature information and generate geological feature identification information, which can perform detailed analysis of geological features and provide a basis for geological situation analysis. However, this method performs feature identification on geological structure data, stratigraphic data, and oil and gas resource data to provide data information for oil and gas exploration. In addition, this method determines the category label of geological information by artificial means, and human factors have a great influence on the classification results of geological features, which is not suitable for shield tunneling construction.
[0005] The present invention provides a method for identifying geological features of shield tunneling based on a fuzzy classification algorithm. According to the reaction of the shield machine to the geological features and the change of shield parameters during the shield construction process, the geological features are classified and marked, and the geological features of the cross-section during the shield tunneling process are accurately obtained. The tunneling time can be reasonably arranged and the shield tunneling process can be warned in advance. Summary of the invention
[0006] The primary purpose of the present invention is to provide a method for identifying geological features of shield tunneling based on a fuzzy classification algorithm. According to the response of the shield machine to the geological features and the changes in the shield parameters during the shield construction process, the geological features are classified and marked, and the geological features of the cross-section during the shield tunneling process are accurately obtained, so that the tunneling time can be reasonably arranged and the shield tunneling process can be warned in advance.
[0007] A further object of the present invention is to provide a shield tunneling geological feature recognition system based on a fuzzy classification algorithm.
[0008] In order to solve the above technical problems, the technical solution of the present invention is as follows:
[0009] A method for identifying geological features of shield tunneling based on fuzzy classification algorithm comprises the following steps:
[0010] S1: Collect geological data, geological feature identification parameters and excavation conditions of the tunnel construction area, wherein the geological data includes geological profiles, and obtain preliminary geological characteristics based on the excavation conditions;
[0011] S2: extracting features of the geological feature identification parameters using a principal component analysis method;
[0012] S3: The data set obtained after feature extraction uses the fuzzy partition coefficient FPC, the silhouette coefficient S and the geological profile to determine the final number of geological feature categories K;
[0013] S4: updating the geological features according to the final category number K of the determined geological features and in combination with the geological profile;
[0014] S5: using the fuzzy classification algorithm, determine the geological feature category to which each shield tunneling ring belongs, and compare it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches above the threshold, the recognition model can be used for geological feature recognition; otherwise, adjust the parameters of the fuzzy classification algorithm until the requirements are met;
[0015] S6: After the newly collected real-time shield parameters are processed in step S2, they are input into the adjusted geological feature recognition model to determine the geological feature category where the shield tunneling is located and output the membership degree of the geological feature category.
[0016] Preferably, the geological profile in step S1 is drawn by a geological survey unit after conducting a soil sampling test on the tunnel section before tunnel construction. The geological feature identification parameters include shield machine parameters and secondary transformation parameters collected and fed back by the shield machine sensor according to the shield construction ring number. The excavated conditions refer to the slag discharged by the screw machine after shield excavation, including the proportion of rock.
[0017] Preferably, the shield machine parameters include shield machine thrust F, propulsion speed AR, cutter head torque T, cutter head speed CRS, upper soil bin pressure UEP and lower soil bin pressure LEP, and the secondary transformation parameters include penetration PR, field cutting depth index FPI, torque cutting depth index TPI and specific energy SE, which are obtained by secondary transformation of shield machine parameters and determined by the following formula:
[0018] PR=AR / CRS
[0019] FPI=F / PR
[0020] TPI=T / PR
[0021]
[0022] Where, AR is the propulsion speed, mm / s; CRS is the cutter head speed, r / s; PR is the penetration, mm / r; FPI is the field cutting depth index, kN·r / mm; TPI is the torque cutting depth index, kN·r; F is the shield thrust, kN; T is the cutter head torque, kN·mm; A is the cutter head area, m 2 ; SE is specific energy, MJ / m 3 .
[0023] Preferably, according to the proportion of the rock, the strata corresponding to the ring number are recorded as:
[0024] When the rock proportion is less than 20%, it is recorded as soft soil formation;
[0025] When the rock proportion is between 20% and 80%, it is recorded as unevenly hard and soft strata;
[0026] When the rock proportion is greater than 80%, it is recorded as a hard rock layer.
[0027] Preferably, in step S2, principal component analysis is used to extract features of the geological feature identification parameters, specifically:
[0028] The principal component analysis method is used to process the n×m geological identification parameters, and the first k principal components are extracted to form a new geological feature identification data set Y, where m is the number of geological feature identification parameters and n is the number of geological parameters recorded according to the ring number.
[0029] Preferably, the principal component analysis specifically comprises the following steps:
[0030] S2.1: Construct an n×m-dimensional recognition parameter database. The database matrix X is represented by the following formula:
[0031]
[0032] In the formula, C m is the name of the mth geological feature identification parameter; X n is the nth row of data in the mth identification parameter; x ij is the value of the j-th row of data for the i-th geological feature identification parameter;
[0033] S2.2: Standardize the database matrix X, and the standardized data X in the i-th row and j-th column ij It is expressed by the following formula:
[0034]
[0035] In the formula, is the average value of the data in the mth column, S m is the standard deviation of the data in the mth column;
[0036] S2.3: Calculate the covariance of the standardized database matrix, construct the covariance matrix Cov, the covariance Cov between the jth factor and the kth factor jk Determined by the following formula:
[0037]
[0038] Where, X ik is the data in the i-th row and k-th column of the database matrix after standardization, is the average of the standardized data of the jth factor and the kth factor. According to the covariance between the factors, the covariance matrix is determined by the following formula:
[0039]
[0040] S2.4: Calculate the eigenvalues and corresponding eigenvectors of the covariance matrix, and calculate the weight of each factor based on the eigenvectors;
[0041] S2.5: Arrange the eigenvectors from left to right into a matrix according to the size of the corresponding eigenvalues, and take the first k columns to form a dimension reduction matrix P;
[0042] S2.6: The original data is converted into k-dimensional data by the following formula:
[0043] Y=XP
[0044] Where Y is the geological feature recognition dataset.
[0045] Preferably, the fuzzy classification algorithm in step S3 is specifically:
[0046] S301: Input the number of classification categories C, weight index m and initial membership function U 0 ;
[0047] S302: Calculate cluster centers (v i ), the cluster center is determined by the following formula:
[0048]
[0049] In the formula, u ik is the membership function, 0 ik ≤1; N is the number of samples; C is the number of classification categories, 1≤k≤N; 1≤i≤C; m is the weight index; v i is the cluster center of the i-th type; x k is the kth sample in the i-th category;
[0050] S303: Update the membership matrix according to the following formula:
[0051]
[0052] Where, d ik =||x k -v i || is the Euclidean distance between the kth sample in the i-th class and the cluster center of the i-th class;
[0053] S304: Given a small value ε>0, determine whether If the conditions are met, the operation process ends and the cluster center, the samples of each class and their membership are output; otherwise, the process returns to S302.
[0054] Preferably, the fuzzy division coefficient in step S3 is a criterion for judging the degree of fuzziness of the fuzzy classification. The clearer the classification, the larger the fuzzy division coefficient. The fuzzy division coefficient is determined by the following formula:
[0055]
[0056] Preferably, the silhouette coefficient S in step S3 is a method for determining the final number of categories K in the fuzzy classification algorithm based on the minimum intra-group distance and the maximum inter-group distance of the sample points in the classification category, and is determined by the following formula:
[0057]
[0058] b i =min{bi1 ,b i2 ,b i3 ,…,b ik}
[0059] In the formula, S i is the silhouette coefficient of a single sample point; a i is the average distance from the i-th sample point to other sample points in the same category; b ik is the average distance from the i-th sample point to all sample points in other categories, and the silhouette coefficient (S) is the average of the silhouette coefficients of all sample points.
[0060] Preferably, the final number of categories K refers to the number of categories K when the fuzzy partition coefficient is the largest and the silhouette coefficient is large; when the maximum value appears when the fuzzy partition coefficient is different from the silhouette coefficient, the geological profile can be used in combination with the two methods to determine the final K value.
[0061] Preferably, the names of the geological features are named according to the soft and hard characteristics of the strata within the shield excavation section in the geological profile. The geological feature names can be named: soft soil strata, uneven soft and hard strata, hard rock strata, soft soil-dominated strata, hard rock-dominated strata, isolated rock strata, etc.
[0062] A shield tunneling geological feature recognition system based on fuzzy classification algorithm, including:
[0063] A collection module, wherein the collection module collects geological data, geological feature identification parameters and excavation conditions of the tunnel construction section, wherein the geological data includes a geological profile, and preliminary geological features are obtained based on the excavation conditions;
[0064] A feature extraction module, wherein the feature extraction module uses a principal component analysis method to extract features of the geological feature identification parameters;
[0065] A category number determination module, wherein the category number determination module uses a fuzzy partition coefficient FPC, a contour coefficient S and a geological profile to determine the final category number K of geological features using the data set obtained after feature extraction;
[0066] An updating module, wherein the updating module updates the geological features according to the determined final category number K of the geological features and in combination with the geological profile;
[0067] A training module, wherein the training module uses a fuzzy classification algorithm to determine the geological feature category to which each shield tunneling ring belongs, and compares it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches a threshold value, the recognition model can be used for geological feature recognition; otherwise, the parameters of the fuzzy classification algorithm are adjusted until the requirements are met;
[0068] The identification module inputs the newly collected real-time shield parameters processed in step S2 into the adjusted geological feature identification model, determines the geological feature category where the shield tunneling is located, and outputs the membership degree of the geological feature category.
[0069] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:
[0070] The present invention utilizes the real-time tunneling parameters of the shield machine, utilizes principal component analysis to reduce the dimension of the data, utilizes fuzzy partition coefficients and contour coefficients to determine the final geological feature categories, and finally utilizes a fuzzy classification algorithm to achieve real-time determination of geological features during the shield tunneling process. On-site engineers can set corresponding shield tunneling parameters, arrange tasks such as tool replacement, etc. according to the geological features determined in real time; in addition, the method can accurately identify various geological states. If abnormal geological features are encountered, the construction process can be warned according to the identification results to ensure the safety of the shield construction. At the same time, by adjusting the geological feature categories and the parameters of the fuzzy classification algorithm, the method can achieve a geological feature recognition accuracy rate of more than 90%, meeting the construction requirements of the shield tunnel. The method is simple and easy to operate, low in cost, can significantly improve the construction efficiency of the shield machine, and ensure the safety of the shield tunneling. BRIEF DESCRIPTION OF THE DRAWINGS
[0071] Figure 1 It is a schematic diagram of the method flow of the present invention.
[0072] Figure 2 A schematic diagram of the changes in the fuzzy partition coefficient and the silhouette coefficient provided in the embodiment.
[0073] Figure 3 A 2D feature space diagram of the geological feature recognition results provided in the embodiment.
[0074] Figure 4 A comparison chart of geological feature identification results and shield construction efficiency provided in the embodiment.
[0075] Figure 5 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION
[0076] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;
[0077] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;
[0078] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.
[0079] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0080] Example 1
[0081] This embodiment provides a shield tunneling geological feature recognition method based on a fuzzy classification algorithm. Figure 1 As shown, the following steps are included:
[0082] S1: Collect geological data, geological feature identification parameters and excavation conditions of the tunnel construction area, wherein the geological data includes geological profiles, and obtain preliminary geological characteristics based on the excavation conditions;
[0083] S2: extracting features of the geological feature identification parameters using a principal component analysis method;
[0084] S3: The data set obtained after feature extraction uses the fuzzy partition coefficient FPC, the silhouette coefficient S and the geological profile to determine the final number of geological feature categories K;
[0085] S4: updating the geological features according to the final category number K of the determined geological features and in combination with the geological profile;
[0086] S5: using the fuzzy classification algorithm, determine the geological feature category to which each shield tunneling ring belongs, and compare it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches above the threshold, the recognition model can be used for geological feature recognition; otherwise, adjust the parameters of the fuzzy classification algorithm until the requirements are met;
[0087] S6: After the newly collected real-time shield parameters are processed in step S2, they are input into the adjusted geological feature recognition model to determine the geological feature category where the shield tunneling is located and output the membership degree of the geological feature category.
[0088] The geological profile in step S1 is drawn by a geological survey unit after a soil sampling test is conducted on the tunnel section before tunnel construction. The geological feature identification parameters include shield machine parameters and secondary transformation parameters collected and fed back by the shield machine sensor according to the shield construction ring number. The excavated conditions refer to the slag discharged by the screw machine after shield excavation, including the proportion of rock.
[0089] The shield machine parameters include shield machine thrust F, propulsion speed AR, cutter head torque T, cutter head speed CRS, upper soil bin pressure UEP and lower soil bin pressure LEP. The secondary transformation parameters include penetration PR, field cutting depth index FPI, torque cutting depth index TPI and specific energy SE, which are obtained by secondary transformation of shield machine parameters and determined by the following formula:
[0090] PR=AR / CRS
[0091] FPI=F / PR
[0092] TPI=T / PR
[0093]
[0094] Where, AR is the propulsion speed, mm / s; CRS is the cutter head speed, r / s; PR is the penetration, mm / r; FPI is the field cutting depth index, kN·r / mm; TPI is the torque cutting depth index, kN·r; F is the shield thrust, kN; T is the cutter head torque, kN·mm; A is the cutter head area, m 2 ; SE is specific energy, MJ / m 3 .
[0095] According to the proportion of rocks, the strata corresponding to the ring numbers are recorded as follows:
[0096] When the rock proportion is less than 20%, it is recorded as soft soil formation;
[0097] When the rock ratio is between 20% and 80%, it is recorded as unevenly hard and soft strata;
[0098] When the rock proportion is greater than 80%, it is recorded as a hard rock layer.
[0099] In step S2, principal component analysis is used to extract features of the geological feature identification parameters, specifically:
[0100] The principal component analysis method is used to process the n×m geological identification parameters, and the first k principal components are extracted to form a new geological feature identification data set Y, where m is the number of geological feature identification parameters and n is the number of geological parameters recorded according to the ring number.
[0101] The principal component analysis specifically comprises the following steps:
[0102] S2.1: Construct an n×m-dimensional recognition parameter database. The database matrix X is represented by the following formula:
[0103]
[0104] In the formula, C m is the name of the mth geological feature identification parameter; X n is the nth row of data in the mth identification parameter; x ij is the value of the j-th row of data for the i-th geological feature identification parameter;
[0105] S2.2: Standardize the database matrix X, and the standardized data X in the i-th row and j-th column ij It is expressed by the following formula:
[0106]
[0107] In the formula, is the average value of the data in the mth column, S m is the standard deviation of the data in the mth column;
[0108] S2.3: Calculate the covariance of the standardized database matrix, construct the covariance matrix Cov, the covariance Cov between the jth factor and the kth factor jk Determined by the following formula:
[0109]
[0110] Where, X ik is the data in the i-th row and k-th column of the database matrix after standardization, is the average of the standardized data of the jth factor and the kth factor. According to the covariance between the factors, the covariance matrix is determined by the following formula:
[0111]
[0112] S2.4: Calculate the eigenvalues and corresponding eigenvectors of the covariance matrix, and calculate the weight of each factor based on the eigenvectors;
[0113] S2.5: Arrange the eigenvectors from left to right into a matrix according to the size of the corresponding eigenvalues, and take the first k columns to form a dimension reduction matrix P;
[0114] S2.6: The original data is converted into k-dimensional data by the following formula:
[0115] Y=XP
[0116] Where Y is the geological feature recognition dataset.
[0117] The fuzzy classification algorithm described in step S3 is specifically:
[0118] S301: Input the number of classification categories C, weight index m and initial membership function U 0 , initial membership function U 0 Determined by the output recognition data set Y and the membership function expression in step S303;
[0119] S302: Calculate cluster centers (v i ), the cluster center is determined by the following formula:
[0120]
[0121] In the formula, u ik is the membership function, 0 ik ≤1; N is the number of samples; C is the number of classification categories, 1≤k≤N; 1≤i≤C; m is the weight index; v i is the cluster center of the i-th type; x k is the kth sample in the i-th category;
[0122] S303: Update the membership matrix according to the following formula:
[0123]
[0124] Where, d ik =||x k -v i || is the Euclidean distance between the kth sample in the i-th class and the cluster center of the i-th class;
[0125] S304: Given a small value ε>0, determine whether If the conditions are met, the operation process ends and the cluster center, the samples of each class and their membership are output; otherwise, the process returns to S302.
[0126] The fuzzy division coefficient in step S3 is a criterion for judging the degree of fuzziness of fuzzy classification. The clearer the classification, the larger the fuzzy division coefficient. The fuzzy division coefficient is determined by the following formula:
[0127]
[0128] The silhouette coefficient S in step S3 is a method for determining the final number of categories K in the fuzzy classification algorithm based on the minimum intra-group distance and the maximum inter-group distance of the sample points in the classification category, and is determined by the following formula:
[0129]
[0130] b i =min{b i1 ,b i2 ,b i3 ,…,b ik}
[0131] In the formula, S i is the silhouette coefficient of a single sample point; a i is the average distance from the i-th sample point to other sample points in the same category; b ik is the average distance from the i-th sample point to all sample points in other categories, and the silhouette coefficient (S) is the average of the silhouette coefficients of all sample points.
[0132] Example 2
[0133] This embodiment provides a specific embodiment of embodiment 1, specifically:
[0134] A double-track shield tunnel is located in Guangzhou. The total length of the tunnel is 3.1 km. The shield tunnel segment is designed to have an inner diameter of 8.0 m, an outer diameter of 8.8 m, and a segment ring width of 1.8 m. Two earth pressure balance shields are used for construction, with a cutter head diameter of 9.15 m. The shield machine mainly passes through fully weathered granite, moderately weathered granite, slightly weathered granite and their mixed strata. Since the tunnel is built in the urban area, a geological survey hole soil sampling test is used to obtain the geological conditions of the shield crossing, and a geological profile is drawn. However, due to the large spacing between the two survey holes, the geological conditions between the survey holes are inferred from the survey hole connection line. The geological conditions are unclear and it is impossible to accurately guide the shield construction. This project adopts a real-time identification method of geological characteristics during shield excavation using a fuzzy classification algorithm for construction.
[0135] In step S2, the collected shield parameters and secondary transformation parameters, totaling 10×1197 pieces of data, are subjected to principal component analysis, and the first k principal components are extracted to form a new geological feature recognition data set Y. The database matrix X is represented by the following formula:
[0136]
[0137] The covariance matrix is:
[0138]
[0139] Calculate the eigenvalues and corresponding eigenvectors of the covariance matrix, and calculate the weight of each factor based on the eigenvector. Arrange the principal components from large to small according to the size of the eigenvector, calculate the cumulative weight of the eigenvector, and the cumulative weight of the first 6 principal components reaches more than 99%, so the 10 identification parameters can be converted into 6 principal components.
[0140] Arrange the eigenvectors from left to right into a matrix according to the size of the corresponding eigenvalues, and take the first 6 columns to form a dimensionality reduction matrix (P). The determined dimensionality reduction matrix is as follows:
[0141]
[0142] The fuzzy classification algorithm is specifically:
[0143] Step 1: Use the traversal method to set the number of classification categories C to 2 to 8, set the weight index m to 1.5, and determine the initial membership function U from the input recognition data set Y 0 .
[0144] Step 2: Calculate the cluster center v i , the cluster center is determined by the following formula:
[0145]
[0146] In the formula, u ikis the membership function, 0 ik ≤1; N is the number of samples (N=1197); C is the number of classification categories, 1≤k≤N; 1≤i≤C; m is the weight index; v i is the cluster center of the i-th type; x k is the kth sample in the i-th class.
[0147] Step 3: Update the membership matrix according to the following formula:
[0148]
[0149] Where, d ik =||x k -v i || is the Euclidean distance between the kth sample in the i-th class and the cluster center of the i-th class.
[0150] Step 4: Given a smaller value (ε>0.0001), determine whether If the conditions are met, the operation process ends and the cluster center, the samples of each category and their membership are output; otherwise, return to step 2.
[0151] (2) Figure 2 The relationship between the fuzzy partition coefficient, the silhouette coefficient and the number of classification categories is given. The fuzzy partition coefficient is determined by the following formula:
[0152]
[0153] (3) According to the minimum intra-group distance and maximum inter-group distance of sample points in the classification category, the silhouette coefficient (S) is determined by the following formula:
[0154]
[0155] b i =min{b i1 ,b i2 ,b i3 ,…,b ik}
[0156] In the formula, S i is the silhouette coefficient of a single sample point; a i is the average distance from the i-th sample point to other sample points in the same category; b ik is the average distance from the ith sample point to all sample points in other categories. The silhouette coefficient (S) is the average of the silhouette coefficients of all sample points.
[0157] (4) Based on Figure 2 As shown in the figure, when the number of classification categories is 3, the fuzzy division coefficient is the largest and the contour coefficient is the largest; combined with the geological profile, the final number of geological feature categories K is determined to be 3.
[0158] According to the final number of geological feature categories (K) determined, the names of geological features are updated in combination with the geological profile. Specifically, the names of geological features are updated as follows: soft soil layer, uneven soft and hard layer, and hard rock layer. The soft soil layer is assigned a value of 1 and named category 1; the uneven soft and hard layer is assigned a value of 2 and named category 2; and the hard rock layer is assigned a value of 3 and named category 3.
[0159] The final number of geological characteristics categories (K), weight index (m), initial membership function (U 0 ) and the reduced dimension identification data set (Y) are input into the fuzzy classification algorithm to determine the geological feature category of each shield tunneling ring, and compared with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature identification. According to the final number of categories (K = 3) and the weight index (m = 1.5), the geological feature type of each ring is determined, such as Figure 3 As shown, a 2D feature space diagram of the geological feature recognition results is drawn based on principal component 1 and principal component 2. The recognition results are compared with the recorded geological feature types, and the accuracy reaches 91.6%, which meets the requirements of shield tunnel construction.
[0160] like Figure 4 As shown, the newly collected real-time parameters of the shield are input into the adjusted geological feature recognition model after the dimensionality reduction processing in the second step, the geological feature category where the shield tunneling is located is determined and the membership degree of the geological feature category is output. Figure 4 The type of geological features in the construction is compared with the construction efficiency. In soft soil, the shield tunneling speed is faster; in hard rock, the shield tunneling speed is the slowest. Figure 4 The construction status is evaluated based on the identification results and the construction status is arranged.
[0161] Example 3
[0162] This embodiment provides a shield tunneling geological feature recognition system based on fuzzy classification algorithm. Figure 5 As shown, including:
[0163] A collection module, wherein the collection module collects geological data, geological feature identification parameters and excavation conditions of the tunnel construction section, wherein the geological data includes a geological profile, and preliminary geological features are obtained based on the excavation conditions;
[0164] A feature extraction module, wherein the feature extraction module uses a principal component analysis method to extract features of the geological feature identification parameters;
[0165] A category number determination module, wherein the category number determination module uses a fuzzy partition coefficient FPC, a contour coefficient S and a geological profile to determine the final category number K of geological features using the data set obtained after feature extraction;
[0166] An updating module, wherein the updating module updates the geological features according to the determined final category number K of the geological features and in combination with the geological profile;
[0167] A training module, wherein the training module uses a fuzzy classification algorithm to determine the geological feature category to which each shield tunneling ring belongs, and compares it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches a threshold value, the recognition model can be used for geological feature recognition; otherwise, the parameters of the fuzzy classification algorithm are adjusted until the requirements are met;
[0168] The identification module inputs the newly collected real-time shield parameters processed in step S2 into the adjusted geological feature identification model, determines the geological feature category where the shield tunneling is located, and outputs the membership degree of the geological feature category.
[0169] The same or similar reference numerals correspond to the same or similar components;
[0170] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;
[0171] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.
Claims
1. A method for identifying geological features of shield tunneling based on fuzzy classification algorithm, characterized in that: The following steps are involved: S1: Collect geological data, geological feature identification parameters and excavation conditions of the tunnel construction area, wherein the geological data includes geological profiles, and obtain preliminary geological characteristics based on the excavation conditions; S2: extracting features of the geological feature identification parameters using a principal component analysis method; S3: The data set obtained after feature extraction uses the fuzzy partition coefficient FPC, the silhouette coefficient S and the geological profile to determine the final number of geological feature categories K; S4: updating the geological features according to the final category number K of the determined geological features and in combination with the geological profile; S5: Using the fuzzy classification algorithm, determine the geological feature category to which each shield tunneling ring belongs, and compare it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches above the threshold, the recognition model can be used for geological feature recognition; otherwise, adjust the parameters of the fuzzy classification algorithm until the requirements are met; S6: after processing in step S2, the newly collected real-time shield parameters are input into the adjusted geological feature recognition model, the geological feature category where the shield tunneling is located is determined, and the membership degree of the geological feature category to which it belongs is output; Step S3 is specifically as follows: S301: Input the number of classification categories C, weight index m and initial membership function U 0 ; S302: Calculate cluster centers (v i ), the cluster center is determined by the following formula: In the formula, u ik is the membership function, 0 ik ≤1; N is the number of samples; C is the number of classification categories, 1≤k≤N; 1≤i≤C; m is the weight index; v i is the cluster center of the i-th type; x k is the kth sample in the i-th category; S303: Update the membership matrix according to the following formula: Where, d ik =||x k -v i || is the Euclidean distance between the kth sample in the i-th class and the cluster center of the i-th class; S304: Given a small value ε>0, determine whether If the conditions are met, the operation process ends and the cluster center, the samples of each class and their membership are output; otherwise, the process returns to S302; The fuzzy division coefficient in step S3 is a criterion for judging the degree of fuzziness of fuzzy classification. The clearer the classification, the larger the fuzzy division coefficient. The fuzzy division coefficient is determined by the following formula: The silhouette coefficient S in step S3 is a method for determining the final number of categories K in the fuzzy classification algorithm based on the minimum intra-group distance and the maximum inter-group distance of the sample points in the classification category, and is determined by the following formula: b i =min{b i1 ,b i2 ,b i3 ,···,b ik } In the formula, S i is the silhouette coefficient of a single sample point; a i is the average distance from the i-th sample point to other sample points in the same category; b ik is the average distance from the i-th sample point to all sample points in other categories, and the silhouette coefficient (S) is the average of the silhouette coefficients of all sample points.
2. The method for identifying geological features of shield tunneling based on fuzzy classification algorithm according to claim 1 is characterized in that: The geological profile in step S1 is drawn by a geological survey unit after a soil sampling test is conducted on the tunnel section before tunnel construction. The geological feature identification parameters include shield machine parameters and secondary transformation parameters collected and fed back by the shield machine sensor according to the shield construction ring number. The excavated conditions refer to the slag discharged by the screw machine after shield excavation, including the proportion of rock.
3. The method for identifying geological features of shield tunneling based on fuzzy classification algorithm according to claim 2 is characterized in that: The shield machine parameters include shield machine thrust F, propulsion speed AR, cutter head torque T, cutter head speed CRS, upper soil bin pressure UEP and lower soil bin pressure LEP. The secondary transformation parameters include penetration PR, field cutting depth index FPI, torque cutting depth index TPI and specific energy SE, which are obtained by secondary transformation of shield machine parameters and determined by the following formula: PR=AR / CRS FPI=F / PR TPI=T / PR Where, AR is the propulsion speed, mm / s; CRS is the cutter head speed, r / s; PR is the penetration, mm / r; FPI is the field cutting depth index, kN·r / mm; TPI is the torque cutting depth index, kN·r; F is the shield thrust, kN; T is the cutter head torque, kN·mm; A is the cutter head area, m 2 ; SE is specific energy, MJ / m 3 .
4. The method for identifying geological features of shield tunneling based on fuzzy classification algorithm according to claim 2 is characterized in that: According to the proportion of rocks, the strata corresponding to the ring numbers are recorded as follows: When the rock proportion is less than 20%, it is recorded as soft soil formation; When the rock proportion is between 20% and 80%, it is recorded as unevenly hard and soft strata; When the rock proportion is greater than 80%, it is recorded as a hard rock layer.
5. The method for identifying geological features of shield tunneling based on fuzzy classification algorithm according to any one of claims 2 to 4, characterized in that: In step S2, principal component analysis is used to extract features of the geological feature identification parameters, specifically: The principal component analysis method is used to process the n×m geological feature identification parameters, and the first k principal components are extracted to form a new geological feature identification data set Y, where m is the number of geological feature identification parameters and n is the number of geological feature identification parameters recorded according to the ring number.
6. The method for identifying geological features of shield tunneling based on fuzzy classification algorithm according to claim 5 is characterized in that: The principal component analysis specifically comprises the following steps: S2.1: Construct an n×m-dimensional recognition parameter database. The database matrix X is represented by the following formula: In the formula, C m is the name of the mth geological feature identification parameter; X n is the nth row of data in the mth identification parameter; x ij is the value of the j-th row of data for the i-th geological feature identification parameter; S2.2: Standardize the database matrix X, and the standardized data X in the i-th row and j-th column ij It is expressed by the following formula: In the formula, is the average value of the data in the mth column, S m is the standard deviation of the data in the mth column; S2.3: Calculate the covariance of the standardized database matrix, construct the covariance matrix Cov, the covariance Cov between the jth factor and the kth factor jk Determined by the following formula: In the formula, X ik is the data in the i-th row and k-th column of the database matrix after standardization, is the average of the standardized data of the jth factor and the kth factor. According to the covariance between the factors, the covariance matrix is determined by the following formula: S2.4: Calculate the eigenvalues and corresponding eigenvectors of the covariance matrix, and calculate the weight of each factor based on the eigenvectors; S2.5: Arrange the eigenvectors from left to right into a matrix according to the size of the corresponding eigenvalues, and take the first k columns to form a dimension reduction matrix P; S2.6: The original data is converted into k-dimensional data by the following formula: Y=XP Where Y is the geological feature recognition dataset.
7. A shield tunneling geological feature recognition system based on fuzzy classification algorithm, characterized in that: The system applies the method described in any one of claims 1 to 5, and the system comprises: A collection module, wherein the collection module collects geological data, geological feature identification parameters and excavation conditions of the tunnel construction section, wherein the geological data includes a geological profile, and preliminary geological features are obtained based on the excavation conditions; A feature extraction module, wherein the feature extraction module uses a principal component analysis method to extract features of the geological feature identification parameters; A category number determination module, wherein the category number determination module uses a fuzzy partition coefficient FPC, a contour coefficient S and a geological profile to determine the final category number K of geological features using the data set obtained after feature extraction; An updating module, wherein the updating module updates the geological features according to the determined final category number K of the geological features and in combination with the geological profile; A training module, wherein the training module uses a fuzzy classification algorithm to determine the geological feature category to which each shield tunneling ring belongs, and compares it with the geological features and geological profiles recorded by the shield tunneling to determine the accuracy of geological feature recognition. If the accuracy reaches a threshold value, the recognition model can be used for geological feature recognition; otherwise, the parameters of the fuzzy classification algorithm are adjusted until the requirements are met; The identification module inputs the newly collected real-time shield parameters processed in step S2 into the adjusted geological feature identification model, determines the geological feature category where the shield tunneling is located, and outputs the membership degree of the geological feature category.
