Real-time distinguishing method and system for tunnel-diameter-spanning surrounding rock category of TBM (Tunnel Boring Machine) based on rock breaking mechanism

Abstract

The invention provides a real-time TBM (Tunnel Boring Machine) cross-tunnel-diameter surrounding rock category distinguishing method and system based on a rock breaking mechanism. The method comprises the following steps: acquiring tunneling data of different TBM construction tunnels, forming a sample data set, extracting tunneling parameters and equipment parameters of each engineering TBM, and forming a refining data set; extracting stable tunneling section data from the refined data set by adopting an MMD method, calculating DFPI and DTPI rock breaking characteristic indexes, and enabling the rock breaking characteristic indexes to correspond to surrounding rock category labels through pile numbers and tunneling time to obtain a surrounding rock grade data set; taking two standardized indexes of DFPI and DTPI as two-dimensional input variables, and constructing a two-dimensional Bayesian surrounding rock classification model; carrying out weighted fusion on [DFPI, DTPI] to obtain an index R, and constructing a one-dimensional surrounding rock discrimination model and a threshold identification method; and inputting the acquired TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the surrounding rock grade of the current TBM tunnel face. According to the method, the model generalization ability is improved, and the problem that a current surrounding rock classification method is insufficient in physical interpretation is solved.

Description

Technical Field

[0001] This invention relates to the field of TBM construction tunnel technology, and in particular to a method and system for real-time identification of surrounding rock type across the tunnel diameter of a TBM based on rock breaking mechanism. Background Technology

[0002] With the continuous advancement of major national infrastructure construction, full-face hardrock tunnel boring machines (TBMs) have been widely used in tunnel projects in my country's railways, water conservancy, power, and municipal engineering sectors. Their excavation efficiency is 3-10 times higher than traditional drill-and-blast methods, and they are suitable for complex tunnel projects with various surrounding rock conditions. Surrounding rock classification is a prerequisite for intelligent tunneling, directly affecting equipment selection and design, cutter layout, parameter settings, and support strategies. It also relates to the early warning and handling of risks such as machine jamming, water inrush, and sudden deformation. However, traditional classification methods rely on sampling and manual interpretation, which lacks continuity and timeliness, making it difficult to meet the real-time control requirements of TBMs. While pure data-driven methods can improve identification accuracy and real-time excavability characterization, they are constrained by the scarcity of tags in the early stages of projects and suffer from domain shifts across equipment and geology, often requiring retraining, thus limiting their widespread application and also facing problems of insufficient interpretability. Summary of the Invention

[0003] In view of this, the present invention provides a method and system for real-time identification of surrounding rock type across tunnel diameter of TBM based on rock breaking mechanism, in order to solve the above problems.

[0004] This invention provides a real-time method for identifying the surrounding rock category of a TBM across tunnel diameters based on rock-breaking mechanisms. The method includes: acquiring tunneling data from different TBM construction tunnels to form a sample dataset; extracting tunneling parameters and equipment parameters of each TBM from the sample dataset to form a refined dataset; using the MMD method to extract stable tunneling section data from the refined dataset, calculating the DFPI and DTPI rock-breaking characteristic indices, and mapping the rock-breaking characteristic indices to surrounding rock category labels using station numbers and tunneling time to obtain a surrounding rock grade dataset; constructing a two-dimensional Bayesian surrounding rock classification model using the standardized DFPI and DTPI as two-dimensional input variables; weighting and fusing [DFPI, DTPI] into an index R to construct a one-dimensional surrounding rock discrimination model and a threshold identification method; and inputting the acquired TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the current surrounding rock grade of the TBM tunnel face.

[0005] In another implementation of the present invention, the step of extracting tunneling parameters and equipment parameters of each engineering TBM from the sample dataset to form a refined dataset includes: extracting tunneling parameters and equipment parameters of each engineering TBM from the sample dataset using a binary discriminant method, and performing data preprocessing to form a refined dataset; the data preprocessing includes calculating the distribution difference of tunneling data using the maximum mean difference and dynamically calculating the sliding window length; for each target point, selecting datasets forward and backward with a fixed sliding window length, calculating the MMD index between the two datasets as a measure of the data distribution difference at that moment in the tunneling cycle segment; and extracting the tunneling parameters using the mean of the MMD index of the entire tunneling segment as a judgment threshold.

[0006] In another implementation of the present invention, the tunneling parameters include the cutterhead rotation speed, tunneling speed, total thrust, cutterhead torque, and penetration depth; the equipment parameters include the diameter of each engineering TBM and the number of cutters.

[0007] In another implementation of the present invention, the formula for calculating the MMD index is:

[0008] Where |X1| and |X2| are the sample numbers of X1 and X2, respectively, and k(x,y) is the kernel function.

[0009] In another implementation of the present invention, the formula for calculating the rock-breaking characteristic index DFPI is:

[0010] Wherein, DFPI is the average thrust penetration index; The effective axial resistance of the cutterhead; N is the number of cutters; For the friction of the shield body; To provide subsequent towing capacity.

[0011] In another implementation of the present invention, the formula for calculating the rock-breaking characteristic index DTPI is:

[0012] Wherein, DTPI is the average torque penetration index; To provide effective rock-breaking resistance torque; This is the resistance torque of the shovel teeth; For mechanical friction resistance; is the inertial torque; D is the diameter of the cutter head.

[0013] In another implementation of the present invention, the formula for calculating the index R is:

[0014] in, This is a weighted synthesis model; w represents the synthesis weights. , For DFPI and DTPI; , denoted as the standard deviation of DFPI and DTPI in the training set.

[0015] Another aspect of the present invention provides a real-time rock classification system for TBM tunnels across diameters based on rock-breaking mechanisms, comprising: a data preprocessing module: acquiring tunneling data of tunnels under construction by different TBMs to form a sample dataset; extracting tunneling parameters and equipment parameters of each TBM project from the sample dataset to form a refined dataset; a feature index calculation module: extracting stable tunneling section data from the refined dataset using the MMD method, calculating DFPI and DTPI rock-breaking feature indices, and mapping the rock-breaking feature indices to rock classification labels using station number and tunneling time to obtain a rock classification dataset; a rock classification model construction module: constructing a two-dimensional Bayesian rock classification model using two standardized indices, DFPI and DTPI, as two-dimensional input variables; weighting and fusing [DFPI, DTPI] into an index R to construct a one-dimensional rock classification model and a threshold identification method; and a real-time rock classification identification module: inputting the acquired TBM tunneling data into the one-dimensional rock classification model to obtain the current rock classification of the TBM tunnel face.

[0016] In another aspect, the present invention provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a real-time TBM cross-diameter surrounding rock classification method based on rock breaking mechanism as described in any of the preceding claims. In another aspect, the present invention provides a computer storage medium storing a computer program, which, when executed by a processor, implements the steps of a real-time TBM cross-diameter surrounding rock classification method based on rock breaking mechanism as described in any of the preceding claims.

[0017] This invention presents a real-time TBM (Tube Breaker) cross-diameter surrounding rock classification method based on rock breaking mechanism. Starting from the rock breaking mechanism and mechanical constraints, it constructs mechanically based mechanistic features (such as DFPI / DTPI) with energy consumption-related quantities such as thrust-torque-penetration as the core to reduce the impact of equipment size and operational disturbances and improve the transferability across tunnel diameters and projects. Furthermore, it introduces an integrated online identification framework based on Bayesian discrimination and threshold rules, combined with MMD segmentation and thrust correction, to achieve continuous, traceable, and updatable determination of surrounding rock classification, thereby providing a real-time and robust foundation for TBM parameter optimization and risk collaborative decision-making. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. By reading the detailed description of the embodiments below, the advantages and benefits of the solutions will become clear to those skilled in the art. The accompanying drawings are only for illustrating preferred embodiments and are not intended to limit the present invention. In the accompanying drawings: Figure 1 This is a schematic diagram of a real-time method for determining the surrounding rock type across the tunnel diameter of a TBM based on rock-breaking mechanism, according to an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of outlier analysis according to an embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram of an MMD partitioning method according to an embodiment of the present invention.

[0021] Figure 4 This is a box-shaped distribution diagram of the average thrust penetration index (DFPI) for different projects according to an embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of the box-type distribution of the average torque penetration index (DTPI) for different engineering processes according to an embodiment of the present invention.

[0023] Figure 6 This is a radar diagram illustrating the separability of surrounding rock categories according to an embodiment of the present invention.

[0024] Figure 7 This is a schematic diagram of the surrounding rock classification decision boundary in the DFPI–DTPI space according to an embodiment of the present invention.

[0025] Figure 8 This is a schematic diagram illustrating the classification confusion matrix of R on XE-VIII, according to an embodiment of the present invention.

[0026] Figure 9 This is a schematic diagram illustrating the classification confusion matrix of R on YCJL, which is an embodiment of the present invention.

[0027] Figure 10 This is a structural block diagram of a real-time TBM (Tunnel Boring Machine) cross-diameter surrounding rock category discrimination system based on rock breaking mechanism, according to an embodiment of the present invention.

[0028] Figure 11 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art should fall within the protection scope of the present invention.

[0030] Figure 1 A schematic diagram of a real-time TBM (Tube Breaker) cross-diameter surrounding rock classification method based on rock breaking mechanism is provided in this embodiment of the invention. Figure 1 As shown, this embodiment mainly includes: S101. Obtain tunneling data for tunnels constructed using different TBMs to form a sample dataset.

[0031] S102. Extract tunneling parameters and equipment parameters of each TBM from the sample dataset to form a refined dataset.

[0032] For example, the refined feature set is effectively extracted, useless data and outliers in the tunneling data are removed, and the MMD method is used to extract stable tunneling section data.

[0033] S103. The MMD method is used to extract stable tunneling section data from the refined dataset, calculate DFPI and DTPI rock breaking characteristic indices, and match the rock breaking characteristic indices with the surrounding rock category labels through station number and tunneling time to obtain the surrounding rock grade dataset.

[0034] For example, the rock breaking characteristic index is obtained according to the calculation methods of DFPI and DTPI. The rock breaking characteristic index is matched with the surrounding rock category label by the station number and tunneling time to obtain the surrounding rock grade dataset. The distribution pattern of key tunneling indicators under ZX, XE-VIII, ABH, GLGS, and KS-VII engineering and surrounding rock conditions is systematically analyzed by box plot method, and Cliff's delta is used to compare the distinguishability of each indicator between different categories.

[0035] S104. Using the two standardized indices DFPI and DTPI as two-dimensional input variables, construct a two-dimensional Bayesian surrounding rock classification model.

[0036] For example, based on this, construction stability data of four typical TBM projects, namely GLGS, KS-VII, ZX and ABH, were selected, and a two-dimensional Bayesian surrounding rock classification model was constructed using two standardized indicators, DFPI and DTPI, as two-dimensional input variables.

[0037] S105. Weigh and fuse [DFPI, DTPI] into index R to construct a one-dimensional surrounding rock discrimination model and threshold identification method.

[0038] For example, [DFPI, DTPI] are weighted and fused into index R to construct a one-dimensional surrounding rock discrimination model and threshold identification method. The cross-engineering generalization ability of index R is verified in XE-VIII and YCJL. After calculating DFPI and DTPI of new tunneling data, the comprehensive index R can be calculated according to its position in the two-dimensional Bayesian model based on the weight superposition. By referring to the corresponding "R-based cross-tunnel diameter surrounding rock classification table", the category of the surrounding rock at the current tunnel face can be determined.

[0039] S106. Input the collected TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the current surrounding rock grade of the TBM tunnel face.

[0040] For example, by combining the cross-engineering surrounding rock classification feature parameters DFPI and DTPI with multiple TBM construction projects of different diameters, a Bayesian surrounding rock classification model can be constructed. This model can achieve real-time identification of surrounding rock categories across engineering projects. Furthermore, the model is based on the rock-machine interaction mechanism, so there is no need to frequently retrain the model. It is not limited by the number of samples and is simple to deploy in practice. It can provide theoretical support and engineering reference for TBM surrounding rock classification, equipment selection, and construction control, and has good cross-tunnel diameter adaptability and application prospects.

[0041] This invention presents a real-time TBM (Tube Breaker) cross-diameter surrounding rock classification method based on rock breaking mechanism. Starting from the rock breaking mechanism and mechanical constraints, it constructs mechanically based mechanistic features (such as DFPI / DTPI) with energy consumption-related quantities such as thrust-torque-penetration as the core to reduce the impact of equipment size and operational disturbances and improve the transferability across tunnel diameters and projects. Furthermore, it introduces an integrated online identification framework based on Bayesian discrimination and threshold rules, combined with MMD segmentation and thrust correction, to achieve continuous, traceable, and updatable determination of surrounding rock classification, thereby providing a real-time and robust foundation for TBM parameter optimization and risk collaborative decision-making.

[0042] In another implementation of the present invention, the step of extracting tunneling parameters and equipment parameters of each engineering TBM from the sample dataset to form a refined dataset includes: extracting tunneling parameters and equipment parameters of each engineering TBM from the sample dataset using a binary discriminant method, and performing data preprocessing to form a refined dataset; the data preprocessing includes calculating the distribution difference of tunneling data using the maximum mean difference and dynamically calculating the sliding window length; for each target point, selecting datasets forward and backward with a fixed sliding window length, calculating the MMD index between the two datasets as a measure of the data distribution difference at that moment in the tunneling cycle segment; and extracting the tunneling parameters using the mean of the MMD index of the entire tunneling segment as a judgment threshold.

[0043] For example, data preprocessing is performed in the following three steps: 1) Extraction of effective data Observation of the raw data revealed that multiple parameters of the data collected during shutdown or when the data acquisition system was malfunctioning were all zero. Based on this characteristic, a binary discriminant function, as shown below, can be used to identify the valid tunneling section data.

[0044]

[0045]

[0046]

[0047] In the formula, d(x) represents the binary discriminant function. When d(x) > 0, it indicates that the data comes from the normal tunneling state. When d(x) = 0, it indicates that the data comes from the shutdown state or the sensor abnormal state. n, v, T, F, and P represent the cutterhead rotation speed, tunneling speed, cutterhead torque, total thrust, and penetration depth, respectively. The tunneling state of the TBM can be determined by the product of the binary discriminant function d(x) of the five key parameters. When D = 0, it indicates that the data entry belongs to the shutdown stage and should be removed.

[0048] 2) Handling outliers, such as... Figure 2 As shown, it includes: (1) Data missing: Due to reasons such as abnormal PLC equipment or network failure, data acquisition may be intermittent or continuous. For intermittent missing values, the average value of the tunneling cycle can be used to fill the gaps; when continuous missing values ​​appear in the tunneling cycle, the data of that tunneling cycle should be directly removed.

[0049] (2) Sensor anomalies: Due to factors such as mechanical vibration and electromagnetic interference, some sensor data may contain outliers, which may adversely affect the analysis results. Therefore, the Raida criterion (3 sigma criterion) should be used to identify outliers and replace them with the average value of the tunneling cycle.

[0050] (3) Incomplete cycle section: Due to the limitation of the propulsion cylinder stroke, the length of a complete tunneling cycle section should not exceed 1.8m (this varies depending on the equipment). However, due to geological conditions, mechanical failures, and other reasons, the length of some tunneling cycle sections does not meet the minimum arch spacing requirement of 0.3m. At this time, the equipment has not reached a stable tunneling state, and such incomplete tunneling cycle sections need to be deleted.

[0051] Stable segment data extraction (1) Determine the length of the sliding window The maximum mean discrepancy (MMD) is used to calculate the distributional discrepancy of the tunneling data, and the following formula is introduced to realize the dynamic calculation of the sliding window length.

[0052]

[0053] Where Lw is the length of the sliding window to be calculated, and Lt is the duration of the tunneling cycle.

[0054] (2) Calculate the MMD index For each target point, the dataset is selected forward and backward respectively, using a sliding window of fixed length. and Then, the MMD index between the two datasets is calculated using the following formula, which measures the difference in data distribution at that moment in the tunneling cycle. Based on this method, the MMD index is calculated sequentially for each moment of the entire tunneling cycle to obtain the overall change in the data.

[0055] (3) Determine the threshold like Figure 3 As shown in Table 1, the average MMD index of the entire tunneling section is used as the judgment threshold. The specific judgment criteria are as follows: When the MMD index starts to exceed the threshold MMDmean and the sample value at that point is less than the average value of the tunneling cycle section, it means that the tunneling state has entered the rising stage from the empty push stage. The index corresponding to this moment is defined as the starting point of the rising stage. When the MMD index starts to fall below the threshold MMDmean and the sample value at that point is greater than the average value of the tunneling cycle section, it indicates that the tunneling state has changed from the rising stage to the stable tunneling stage. The index corresponding to this moment is defined as the starting point of the stable stage. When the MMD index starts to exceed the threshold MMDmean and the sample value at that point is greater than the average value of the tunneling cycle section, it means that the tunneling state has entered the falling stage from the stable tunneling stage. The index corresponding to this moment is defined as the ending point of the stable stage. Based on this, the stable section data is extracted.

[0056] Table 1 Judgment Criteria for Each Stage of the Cycle

[0057] In another implementation of the present invention, the tunneling parameters include the cutterhead rotation speed, tunneling speed, total thrust, cutterhead torque, and penetration depth; the equipment parameters include the diameter of each engineering TBM and the number of cutters.

[0058] For example, five tunneling parameters—cutterhead rotation speed n, tunneling speed v, total thrust F, cutterhead torque T, and penetration depth P—and two equipment parameters—diameter D and number of cutters N—of the TBM for each project are extracted to form a refined dataset.

[0059] In another implementation of the present invention, the formula for calculating the MMD index is:

[0060] Where |X1| and |X2| are the sample numbers of X1 and X2, respectively, and k(x,y) is the kernel function.

[0061] In another implementation of the present invention, the formula for calculating the rock-breaking characteristic index DFPI is:

[0062] Wherein, DFPI is the average thrust penetration index. ; The effective axial resistance of the cutterhead; N is the number of cutters; For the friction of the shield body; To provide subsequent towing capacity.

[0063] For example, DFPI (Diameter-independent Force Penetration Index) is the average thrust penetration index, which represents the average thrust required for a single cutter to advance a unit distance.

[0064] In another implementation of the present invention, the formula for calculating the rock-breaking characteristic index DTPI is:

[0065] Wherein, DTPI is the average torque penetration index. ; To effectively break rock resistance torque, ; This is the resistance torque of the shovel teeth; For mechanical friction resistance; is the inertial torque; D is the diameter of the cutter head, in meters.

[0066] For example, DTPI (Diameter-independent Torque Penetration Index) is the average torque penetration index, representing the average torque required to tunnel a unit distance. From the perspective of the interaction between rock machine information and the tunneling machine, DFPI and DTPI characterize the tunneling performance of the TBM.

[0067] In another implementation of the present invention, the formula for calculating the index R is:

[0068] in, This is a weighted synthesis model; w represents the synthesis weights. , For DFPI and DTPI; , denoted as the standard deviation of DFPI and DTPI in the training set.

[0069] For example, in actual engineering, monitoring systems typically only record the total cutterhead thrust (F) and cutterhead torque (T), without directly distinguishing their components. Therefore, the no-load thrust and no-load torque are defined as follows:

[0070]

[0071] The unloaded thrust differs from the total thrust and cannot be collected in real time. In this invention, five groups of unloaded tunneling tests, totaling 30 groups, were conducted in six projects for different surrounding rock types. Thrust and torque data of the cutterhead unloaded thrust were collected under stable tunneling conditions to obtain the average unloaded thrust and torque values ​​for each project.

[0072] Furthermore, a dataset of surrounding rock grades is established: A cross-diameter surrounding rock grade dataset was constructed based on six engineering data points: ZX, XE-VIII, ABH, GLGS, KS-VII, and YCJL, totaling 26 km. The surrounding rock labels are associated with the tunneling data through tunneling time and station number. Then, the distribution patterns of key tunneling indicators under engineering and surrounding rock conditions for ZX, XE-VIII, ABH, GLGS, and KS-VII were systematically analyzed using the boxplot method, and Cliff's delta was used to compare the distinguishability of each indicator among different categories.

[0073]

[0074] Where #(x>y) represents all , In pairwise comparisons The number of times (similarly #(x) <y)), , These represent the two sample sizes. δ>0 indicates that X is often greater than Y, and δ<0 indicates the opposite. The larger |δ| is, the easier it is to distinguish between the two distributions. This study first aggregates samples from all projects for each category (after one balanced sampling and IQR to remove extremes), and then for each pair of categories ( , )Calculate δ.

[0075] Based on this, construction stability data from four typical TBM projects (GLGS, KS-VII, ZX, and ABH) were selected. A two-dimensional Bayesian rock classification model was constructed using standardized indices DFPI and DTPI as two-dimensional input variables. The weighted fusion of [DFPI, DTPI] into index R was used to construct a one-dimensional rock discrimination model and threshold identification method. The cross-project generalization ability of index R was verified in XE-VIII and YCJL.

[0076] Furthermore, assuming that each surrounding rock category follows a two-dimensional Gaussian distribution in the DFPI–DTPI space, for each category... Sample vector The class-conditional probability density function is:

[0077] in, Let be the mean vector of category c. Let be the covariance matrix of class c.

[0078] The core of training a multidimensional Bayesian model lies in the accurate estimation of the mean vector and covariance matrix of each class. For each class c, its mean... With covariance The estimation formula is as follows:

[0079] in, This represents the number of samples in the training set that belong to class c. This is the i-th sample in this category.

[0080] According to Bayes' theorem, the posterior probability expression is:

[0081] in, This represents the prior probability, typically the proportion of each class in the training set. The final predicted class is determined by the class corresponding to the highest posterior probability.

[0082] If we further define any two categories , If the discriminant function between them is satisfied, then the decision boundary is determined by satisfying... The points constitute the set, meaning the posterior probabilities are equal:

[0083] Furthermore, [DFPI, DTPI] are weighted and fused into an index R, which simplifies the discrimination process while maintaining accuracy, thereby improving the convenience of engineering applications.

[0084] Depend on Figure 7 The model results can effectively reflect the relative position and classification rules of each category in the feature space. In practical applications, the new tunneling feature parameters DFPI and DTPI are substituted into the two-dimensional Gaussian distribution model to determine the surrounding rock category, thereby realizing the automatic identification of different surrounding rock categories and assisting in on-site judgment.

[0085] Before constructing the comprehensive index, to avoid the interference of differences in measurement units on the results, this paper first standardizes the feature vectors, and then establishes the weighted synthesis model as follows:

[0086] in, , denoted as the standard deviation of DFPI and DTPI in the training set. , For DFPI and DTPI Assume the mean of each class is , Then the posterior mean of all categories is:

[0087] After normalization, it is used as the synthesis weight w:

[0088] Finally, the standardized DFPI and DTPI are combined into an index R.

[0089] To achieve one-dimensional probabilistic modeling of the R-value, this paper fits the R-distribution of each category in the training set using one-dimensional kernel density estimation (KDE). Specifically, for any category c, its KDE estimation expression is:

[0090] in, The number of samples in category c; Kernel function (commonly Gaussian kernel); h: bandwidth parameter; The R-value of the i-th sample in this category; For any sample to be predicted, its corresponding R value can be substituted into various KDE functions to calculate its likelihood value for each category. The predicted category is determined according to the maximum likelihood criterion.

[0091] Furthermore, the likelihood values ​​can be normalized to obtain the posterior probability expressions for each class:

[0092] To construct a one-dimensional classification rule based on the R value, the intersection points between the KDE curves of each category are used as the category discrimination thresholds in this paper. Let the intersection point sequence be { }, then the following interval discrimination rule can be established:

[0093] Finally, after calculating the DFPI and DTPI for the newly excavated data, according to the position in the two-dimensional Bayesian model, the comprehensive index R can be calculated by weighted superposition. By referring to the "Classification Table of Surrounding Rock with Different Borehole Diameters Based on R", the category of the surrounding rock of the current tunnel face can be determined.

[0094] The one-dimensional piecewise discrimination rule table based on the R value is shown in Table 2 below. This rule can be expressed as: when 0.651 < R ≤ 1.270, it is determined as ClassⅣ; when 1.270 < R ≤ 1.661, it is determined as ClassⅢ. This rule table combines the natural intersection points of the probability density curves, does not rely on artificial experience thresholds, and has strong objectivity and adaptability.

[0095] Table 2 Classification Table of Surrounding Rock with Different Borehole Diameters Based on R

[0096] Example 1 Collect the tunneling data of six diameters including ZX, XE-Ⅷ, ABH, GLGS, KS-Ⅶ, and YCJL as shown in the following table, with the range covering 4.03 - 9.03m. The surrounding rock types passed through by each project cover various lithologic combinations such as mudstone, siltstone, dolomite, tuff, shale, and granite. The surrounding rock grading covers gradesⅡ - Ⅴ, and widely involves TBM construction projects in soft rock, hard rock, and composite strata, with a total of 26km of tunneling data, to establish a sample data set.

[0097] Table 3 Main Technical Parameters of TBM [[ID=2…]]

[0098] At the same time of collecting the on-site tunneling data during the approach, the corresponding surrounding rock conditions are also obtained, and the classification grade of the rock is determined according to the HC (Hock - Brown criterion) method. According to the HC moisture method, the surrounding rock is divided into five grades:Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ. ClassⅠ surrounding rock is the hardest and most intact, while ClassⅤ surrounding rock is the weakest and most broken. In the establishment of the data set of this invention, four types of surrounding rock, namelyⅡ,Ⅲ,Ⅳ, andⅤ, are collected. [[ID=3…]]

[0099] Example 2 Based on five engineering data sets (ZX, XE-VIII, ABH, GLGS, and KS-VII), a classification standard for surrounding rock across tunnel diameters was constructed, and the accuracy of the model was verified. The distribution patterns of key tunneling indicators under engineering and surrounding rock conditions were analyzed for ZX, XE-VIII, ABH, GLGS, and KS-VII projects, and box plots were used to visualize each indicator. Figure 4 The data indicates that the DFPI (Digital Degradation Intensity) of different engineering projects decreases with increasing surrounding rock type, with the average value exceeding 50 under Class II surrounding rock conditions. It is clearly distinguishable from other categories; Figure 5 The results show that DTPI decreases significantly with the increase of surrounding rock category, indicating that the tunneling difficulty decreases with the increase of surrounding rock category. DFPI and DTPI further reduce operational disturbances while eliminating the influence of equipment, and can more accurately reflect the tunneling characteristics of surrounding rock. They maintain consistency and focus under different engineering and category combinations, demonstrating good cross-engineering versatility and classification identification ability.

[0100] It prioritizes both online availability and interpretability, with a low deployment threshold: MMD adaptive segmentation enables automatic extraction of stable segments; two-dimensional Bayes provides posterior probabilities and clear decision boundaries; one-dimensional R + KDE intersection thresholds compress the model into a rule table, facilitating rapid deployment and maintenance in low-computing-power environments on-site. The classification criteria are traceable, and the thresholds can be updated on a rolling basis, meeting the real-time and interpretability requirements of engineering projects.

[0101] The data from the five projects were combined according to the surrounding rock category, and Cliff's delta was used to compare the discriminative power of each indicator between different categories:

[0102] Where #(x>y) represents all , In pairwise comparisons The number of times (similarly #(x) <y)), , These represent the two sample sizes. δ>0 indicates that X is often greater than Y, and δ<0 indicates the opposite. The larger |δ| is, the easier it is to distinguish between the two distributions. This study first aggregates samples from all projects for each category (after one balanced sampling and IQR to remove extremes), and then for each pair of categories ( , )Calculate δ.

[0103] Figure 6This indicates that compared to the indicators n, v, P, T, and F, DTPI and DFPI have the strongest discriminative power: most category pairs have a value ≥0.85, with II–V reaching 1.00 and II–IV also at 1.00 / 0.99 respectively. P shows high separability with P in pairs containing "Class II" (II–IV and II–V are both greater than 0.85), but weakens rapidly between adjacent categories. Performance is uneven, reaching only 0.71 in III–V, with most others between 0.22 and 0.40. v and n are generally weak, such as v for III–IV = 0.16 and n for II–III = 0.13, but still at a moderate level in IV–V (approximately 0.51–0.53). Overall, pairs containing Class II are the easiest to distinguish, while pairs with larger adjacent rock types (such as IV–V) are the most difficult to distinguish. The overall discriminative power, from strongest to weakest, is as follows: .

[0104] Mechanism-Data Dual-Driven, Transferable Across Tunnels: Unlike pure data-driven schemes that rely on dense tags, this invention uses DFPI / DTPI as a bridge to directly couple thrust-torque-penetration with rock-breaking power / energy consumption mechanisms. Combined with air thrust calibration and δ index optimization, it achieves stronger generalization capabilities across tunnel diameters, equipment, and engineering projects under physical consistency constraints, significantly reducing reliance on the initial tagless stage and the difficulty of cold start.

[0105] Example 3 To further verify the cross-engineering generalization ability of index R, it was validated in two projects, XE-VIII and YCJL, and the results are as follows. Figure 8 , Figure 9 As shown. Figure 8 The results show that the synthetic index R performs well on the test set (XE-VIII) with close geographical proximity, with an overall accuracy of 90% and an F1 value of 0.85. It is particularly effective in identifying Class II surrounding rocks, indicating that under the condition of consistent data distribution, the index can stably characterize the characteristics of surrounding rock categories. Figure 9 The results of the one-dimensional classification model on another project, YCJL, show that despite differences in geological distribution and training set, the model still achieved an accuracy of 76%. The recognition performance of Class III and Class IV remained good, reflecting that the synthetic index has a certain generalization ability across different projects and has the potential for further application expansion.

[0106] Robustness and improved data governance: Through a complete process of effective segment discrimination, missing / outlier handling, empty push correction and cross-engineering verification, noise and bias are suppressed throughout the entire link from data acquisition to discrimination, ensuring the stability of indicators and the reliability of recognition, which is superior to existing methods that rely solely on offline training and lack a mechanism for verification.

[0107] This application is not limited by surrounding rock labels and data homogeneity, can improve the generalization ability of the model, make up for the lack of physical interpretation of the current surrounding rock classification method, and realize real-time dynamic identification of surrounding rock categories across tunnel diameters.

[0108] Another aspect of the present invention provides a real-time identification system for the surrounding rock category across the tunnel diameter of a TBM based on rock-breaking mechanism, comprising: Data preprocessing module 310: acquires tunneling data of tunnels under different TBM constructions to form a sample dataset; extracts tunneling parameters and equipment parameters of each TBM from the sample dataset to form a refined dataset.

[0109] Feature index calculation module 320: The MMD method is used to extract stable tunneling section data from the refined dataset, calculate DFPI and DTPI rock breaking feature indices, and match the rock breaking feature indices with surrounding rock category labels through station number and tunneling time to obtain surrounding rock grade dataset.

[0110] Module 330 for constructing a surrounding rock classification model: Using two standardized indicators, DFPI and DTPI, as two-dimensional input variables, a two-dimensional Bayesian surrounding rock classification model is constructed; [DFPI, DTPI] are weighted and fused into an indicator R to construct a one-dimensional surrounding rock discrimination model and a threshold identification method.

[0111] Real-time surrounding rock category identification module 340: Inputs the collected TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the current surrounding rock grade of the TBM tunnel face.

[0112] The present invention relates to a real-time TBM surrounding rock classification system based on rock-breaking mechanism. Starting from the rock-breaking mechanism and mechanical constraints, it constructs mechanically based mechanism features (such as DFPI / DTPI) with energy consumption-related quantities such as thrust-torque-penetration as the core to reduce the impact of equipment size and operational disturbances and improve the transferability across tunnel diameters and projects. Furthermore, it introduces an integrated online identification framework based on Bayesian discrimination and threshold rules, combined with MMD segmentation and thrust correction, to achieve continuous, traceable, and updatable determination of surrounding rock classification, thereby providing a real-time and robust foundation for TBM parameter optimization and risk collaborative decision-making.

[0113] Another aspect of the present invention, such as Figure 11As shown, the electronic device includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage section 1008 into a random access memory (RAM) 1003. The RAM 1003 also stores various programs and data required for the system's operating instructions. The CPU 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0114] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. A removable medium 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 1010 as needed so that computer programs read from it can be installed into storage section 1008 as needed.

[0115] Specifically, according to embodiments of this application, the flowchart above refers to... Figure 1 The described process can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such an embodiment, the computer program contains program code for performing the methods shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 1009, and / or installed from removable medium 1011. When the computer program is executed by central processing unit (CPU) 1001, it performs the functions defined in the system of this application.

[0116] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operational instructions of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two connected blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified functions or operational instructions, or using a combination of dedicated hardware and computer instructions.

[0118] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be located in a processor. The names of these units or modules do not, in certain circumstances, constitute a limitation on the unit or module itself.

[0119] An exemplary embodiment of this application also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods of various embodiments of this application.

[0120] The methods described above according to embodiments of the present invention can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored on a remote recording medium or a non-transitory machine-readable medium and subsequently stored on a local recording medium, downloaded via a network. Thus, the methods described herein can be processed by software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., RAM, ROM, flash memory, etc.) capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for executing the methods shown herein.

[0121] Specific embodiments of the present invention have now been described. Other embodiments are within the scope of the appended claims. In some cases, the actions described in the claims can be performed in a different order and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result.

[0122] It should be noted that all directional indications (such as up, down, left, right, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship between the components in a certain order (as shown in the figure). If the specific order changes, the directional indication will also change accordingly.

[0123] In the description of this invention, the terms "first" and "second" are used only for convenience in describing different components or names, and should not be construed as indicating or implying a sequential relationship, relative importance, or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" and "second" may explicitly or implicitly include at least one of that feature.

[0124] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0125] It should be noted that although specific embodiments of the present invention have been described in detail with reference to the accompanying drawings, this should not be construed as limiting the scope of protection of the present invention. Various modifications and variations that can be made by those skilled in the art without inventive effort within the scope described in the claims still fall within the scope of protection of the present invention.

[0126] The examples of the embodiments of the present invention are intended to concisely illustrate the technical features of the embodiments of the present invention, so that those skilled in the art can intuitively understand the technical features of the embodiments of the present invention, and are not intended to be an improper limitation of the embodiments of the present invention.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A TBM cross-hole diameter surrounding rock category real-time discrimination method based on rock breaking mechanism, characterized in that, The method comprises the following steps: obtaining tunneling data of different TBM construction tunnels to form a sample data set; extracting tunneling parameters and equipment parameters of each TBM from the sample data set to form a refined data set; extracting stable tunneling section data from the refined data set by using the MMD method, calculating the DFPI and DTPI rock breaking characteristic indexes, and corresponding the rock breaking characteristic indexes with surrounding rock category labels through the stake number and tunneling time to obtain a surrounding rock grade data set; using the two standardized indexes DFPI and DTPI as two-dimensional input variables to construct a two-dimensional Bayesian surrounding rock classification model; fusing [DFPI, DTPI] into an index R by weighting to construct a one-dimensional surrounding rock discrimination model and a threshold identification method; inputting the collected TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the current TBM tunnel face surrounding rock grade.

2. The method of claim 1, wherein, The step of extracting tunneling parameters and equipment parameters of each TBM from the sample data set to form a refined data set comprises the following steps: extracting tunneling parameters and equipment parameters of each TBM from the sample data set by using a binary discrimination method and performing data preprocessing to form a refined data set; the data preprocessing comprises calculating the distribution difference of the tunneling data by using the maximum mean difference and dynamically calculating the length of the sliding window; for each target point, a sliding window is used as a fixed length, and data sets are selected forward and backward respectively, and the MMD index between the two data sets is calculated as a measure of the data distribution difference at that time in the tunneling cycle section; the mean value of the MMD index of the entire tunneling section is used as a judgment threshold to extract the tunneling parameters.

3. The method of claim 2, wherein, The tunneling parameters include the extraction of the cutter head speed, the tunneling speed, the total thrust, the cutter head torque and the penetration depth. The equipment parameters include the diameter of each TBM and the number of rollers.

4. The method of claim 2, wherein, The calculation formula of the MMD index is as follows: wherein |X1| and |X2| are the sample numbers of X1 and X2 respectively, and k(x, y) is a kernel function.

5. The method of claim 1, wherein, The calculation formula of the rock breaking characteristic index DFPI is as follows: Wherein, DFPI is the average thrust penetration index; is the effective cutterhead axial resistance; N is the number of cutters; is the shield friction force; is the rear matching tractive force.

6. The method of claim 5, wherein, The calculation formula of the rock breaking characteristic index DTPI is as follows: where DTPI is the average penetration index of torque; is the effective rock breaking resistance torque; is the resistance torque of the pick; is the mechanical friction resistance; is the inertia torque; and D is the diameter of the cutter head.

7. The method of claim 6, wherein, The calculation formula of the index R is as follows: wherein, is a weighted synthetic model; w is a synthetic weight; , is the DFPI and DTPI; , is the standard deviation of the DFPI and DTPI in the training set.

8. A TBM cross-hole diameter surrounding rock category real-time discrimination system based on rock breaking mechanism, characterized in that, The method comprises the following steps: a data preprocessing module: obtaining tunneling data of different TBM construction tunnels to form a sample data set; extracting tunneling parameters and equipment parameters of each TBM from the sample data set to form a refined data set; a feature index calculation module: extracting stable tunneling section data from the refined data set by using the MMD method, calculating the DFPI and DTPI rock breaking characteristic indexes, and corresponding the rock breaking characteristic indexes with surrounding rock category labels through the stake number and tunneling time to obtain a surrounding rock grade data set; a surrounding rock classification model construction module: using the two standardized indexes DFPI and DTPI as two-dimensional input variables to construct a two-dimensional Bayesian surrounding rock classification model; fusing [DFPI, DTPI] into an index R by weighting to construct a one-dimensional surrounding rock discrimination model and a threshold identification method; a surrounding rock category real-time identification module: inputting the collected TBM tunneling data into the one-dimensional surrounding rock discrimination model to obtain the current TBM tunnel face surrounding rock grade.

9. An electronic device, comprising: The method comprises the following steps: The memory, the processor and the computer program stored on the memory and capable of running on the processor, wherein the processor implements the steps of the TBM trans-hole-diameter surrounding rock category real-time discrimination method based on rock breaking mechanism according to any one of claims 1 to 7 when executing the computer program.

10. A computer storage medium, characterized in that The computer storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the TBM trans-hole-diameter surrounding rock category real-time discrimination method based on rock breaking mechanism according to any one of claims 1 to 7.

Images