Tunnel surrounding rock quality prediction method and system
By integrating multi-source data features from tunnel boring machines, including rock debris images and vibration signals, the problems of lag and accuracy in traditional surrounding rock quality evaluation have been solved, enabling higher-precision real-time evaluation and construction decision support.
Patent Information
- Application Number
- CN202510982623.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-16
- Publication Date
- 2025-11-18
AI Technical Summary
Traditional methods for evaluating the quality of tunnel surrounding rock are characterized by strong lag, high cost, and difficulty in reflecting the condition of the surrounding rock at the tunnel face in real time. Existing technologies have failed to fully explore the correlation and complementarity between multiple data sources, resulting in insufficient accuracy of evaluation results.
By acquiring multi-source data during the tunnel boring machine's excavation process, including rock debris image data, tunnel boring machine excavation data, and vibration signal data, features are extracted and fused. Pre-trained models are used for prediction, and a decision fusion strategy is combined to obtain the final prediction result of the surrounding rock quality.
It enables a more comprehensive and accurate evaluation of the surrounding rock quality, overcomes the limitations of single-data evaluation, improves evaluation accuracy, meets the real-time decision-making needs of TBM construction, and reduces construction costs.
Smart Images
Figure CN120974885A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of civil engineering, and in particular to a tunnel surrounding rock quality prediction method and system. BACKGROUND
[0002] As a kind of efficient and safe tunnel construction equipment, Full Face Tunnel Boring Machine (TBM) has been widely used in various tunnel projects. In the process of TBM construction, real-time and accurate evaluation of surrounding rock quality plays a decisive role in construction safety, tunneling parameter optimization and support design. Traditional surrounding rock quality evaluation methods mainly rely on geological survey, field sampling and laboratory testing. These methods have strong hysteresis, high cost and difficulty in reflecting the surrounding rock conditions of the tunneling face in real time.
[0003] In recent years, with the development of sensor technology and data acquisition system, a large amount of real-time data can be obtained in the process of TBM construction, including tunneling parameters, rock slag information and vibration signals. These multi-source data reflect the characteristics of surrounding rock from different angles, but existing technologies often use only a single data source for evaluation, or simply superimpose multi-source information, failing to fully exploit the correlation and complementarity between data, resulting in insufficient accuracy of evaluation results. SUMMARY
[0004] In view of the above-mentioned defects or shortcomings in the prior art, it is desirable to provide a tunnel surrounding rock quality prediction method and system, which has the advantages of comprehensive and accurate surrounding rock quality evaluation.
[0005] In a first aspect, the embodiments of the present application provide a tunnel surrounding rock quality prediction method, comprising:
[0006] obtaining actual multi-source data in the process of tunnel boring machine tunneling, wherein the multi-source data at least includes rock slag image data of surrounding rock, tunneling data of tunnel boring machine and vibration signal data;
[0007] extracting rock slag image data features of the rock slag image data, tunneling data features and vibration signal data features, respectively;
[0008] performing feature fusion on the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features, and obtaining the final prediction result of the surrounding rock quality according to the fused features, or respectively obtaining the first prediction result of the surrounding rock quality according to the rock slag image data features, the second prediction result of the surrounding rock quality according to the tunneling data features and the third prediction result of the surrounding rock quality according to the vibration signal data features, and deciding the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
[0009] In some examples, the rock image data features, the tunneling data features, and the vibration signal data features of the rock image data are fused, and a final prediction result of the surrounding rock quality is obtained according to the fused features, including:
[0010] The rock image data features, the tunneling data features, and the vibration signal data features of the rock image data are fused, and a final prediction result of the surrounding rock quality is obtained according to the fused features, including:
[0011] The fused features are input into a pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, wherein the first prediction model is pre-trained with a plurality of multi-source data of a tunnel boring machine tunneling process as training samples and with a surrounding rock quality evaluation result as a label.
[0012] In some examples, before the fused features are input into the pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, the method further includes:
[0013] The plurality of multi-source data of the tunnel boring machine tunneling process;
[0014] Optimizing each type of multi-source data;
[0015] After the optimization, the features of each type of multi-source data are extracted;
[0016] The features of each type of multi-source data are fused to obtain the fused features;
[0017] The initial first prediction model is trained according to the fused features to obtain the pre-trained first prediction model.
[0018] In some examples, the optimization of each type of multi-source data includes:
[0019] Removing outliers in each type of multi-source data;
[0020] After removing the outliers in each type of multi-source data, each type of multi-source data is aligned.
[0021] In some examples, after the optimization, the features of each type of multi-source data are extracted, including:
[0022] The tunneling cycle is divided from the tunneling data, and the tunneling data features are extracted according to the divided tunneling cycle, wherein the tunneling data features include the mean, variance, and range of the tunneling data;
[0023] The rock image data features of the surrounding rock are extracted from the rock image data of the surrounding rock, and the rock image data features of the surrounding rock include the size and shape features of each rock block;
[0024] extracting vibration signal data features from the vibration signal data, wherein the vibration signal data features comprise time domain features and frequency domain features.
[0025] In some examples, the training of the initial first prediction model according to the fusion features to obtain the pre-trained first prediction model comprises:
[0026] inputting the fusion features into the initial first prediction model, and training the initial first prediction model according to a loss before an output of the initial first prediction model and a label until the loss before the output of the initial first prediction model and the label meets a set requirement, to obtain the pre-trained first prediction model.
[0027] In some examples, the obtaining of the first prediction result of the surrounding rock quality according to the surrounding rock image data features, the obtaining of the second prediction result of the surrounding rock quality according to the tunneling data features, and the obtaining of the third prediction result of the surrounding rock quality according to the vibration signal data features, and the decision of the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result comprise:
[0028] inputting the surrounding rock image data features into the pre-trained second prediction model to obtain the first prediction result of the surrounding rock quality;
[0029] inputting the tunneling data features into the pre-trained third prediction model to obtain the second prediction result of the surrounding rock quality;
[0030] inputting the vibration signal data features into the pre-trained fourth prediction model to obtain the third prediction result of the surrounding rock quality;
[0031] decision fusion of the first prediction result, the second prediction result and the third prediction result to determine the final prediction result of the surrounding rock quality.
[0032] In some examples, the decision fusion of the first prediction result, the second prediction result and the third prediction result to determine the final prediction result of the surrounding rock quality comprises:
[0033] voting on the first prediction result, the second prediction result and the third prediction result respectively;
[0034] the prediction result with the most votes is taken as the final prediction result of the surrounding rock quality.
[0035] In some examples, the final prediction result of the surrounding rock quality further comprises:
[0036] Based on the final prediction results of the surrounding rock quality, the construction decision of the tunnel boring machine is adjusted.
[0037] Secondly, embodiments of this application provide a tunnel surrounding rock quality prediction system, including:
[0038] The acquisition module is used to obtain actual multi-source data during the tunnel boring machine's tunneling process. The multi-source data includes at least rock debris image data of the surrounding rock, tunnel boring machine tunneling data, and vibration signal data.
[0039] The extraction module is used to extract the rock debris image data features, tunneling data features, and vibration signal data features from the rock debris image data, respectively.
[0040] The prediction module is used to perform feature fusion on the rock debris image data features, tunneling data features, and vibration signal data features of the rock debris image data, and obtain the final prediction result of the surrounding rock quality based on the fused features. Alternatively, it can obtain a first prediction result of the surrounding rock quality based on the rock debris image data features, a second prediction result of the surrounding rock quality based on the tunneling data features, and a third prediction result of the surrounding rock quality based on the vibration signal data features, and determine the final prediction result of the surrounding rock quality based on the first prediction result, the second prediction result, and the third prediction result.
[0041] Thirdly, embodiments of this application provide a computing device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the tunnel surrounding rock quality prediction method according to the first aspect and any possible method in the first aspect.
[0042] After obtaining actual multi-source data in the tunneling process of the tunnel boring machine in the embodiments of the present application, the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features are extracted, and finally the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features are fused, and the final prediction result of the surrounding rock quality is obtained according to the fused features, or the first prediction result of the surrounding rock quality is obtained according to the rock slag image data features, the second prediction result of the surrounding rock quality is obtained according to the tunneling data features, and the third prediction result of the surrounding rock quality is obtained according to the vibration signal data features, and the final prediction result of the surrounding rock quality is decided according to the first prediction result, the second prediction result and the third prediction result. By reasonably fusing the multi-source information such as TBM tunneling data, rock slag image and vibration signal, more comprehensive and accurate evaluation of the surrounding rock quality can be realized, which has the following advantages: high comprehensive evaluation, fusion of three complementary data sources of TBM tunneling data, rock slag image and vibration signal, overcoming the limitations of single data evaluation; high accuracy, by feature fusion and decision fusion strategies, the associated information between data is fully mined, and the evaluation accuracy is improved; strong real-time, based on real-time data acquisition for rapid analysis, meeting the real-time decision-making needs of TBM construction, strong self-adaptability, intelligent algorithm can adapt to different geological conditions, with strong generalization ability; good engineering practicability, completely based on TBM construction process data, without additional exploration, reducing construction cost.
[0043] Additional aspects and advantages of the application will be set forth in part in the description that follows, and in part will become apparent to those skilled in the art upon examination of the following description and drawings. BRIEF DESCRIPTION OF DRAWINGS
[0044] Other features, objects, and advantages of the application will become more apparent from the following detailed description when read in connection with the following drawings:
[0045] Figure 1 Flow chart of the tunnel surrounding rock quality prediction method of an embodiment of the present application;
[0046] Figure 2 Rock mass identification flow chart of the tunnel surrounding rock quality prediction method of an embodiment of the present application fusing TBM tunneling data, rock slag image and vibration signal;
[0047] Figure 3 Rock slag image segmentation effect schematic diagram of the tunnel surrounding rock quality prediction method of an embodiment of the present application under different surrounding rock grades;
[0048] Figure 4 Schematic diagram of a complete cycle section of TBM tunneling data of the tunnel surrounding rock quality prediction method of an embodiment of the present application;
[0049] Figure 5 A TBM vibration signal information diagram of a tunnel surrounding rock quality prediction method according to an embodiment of the present application is shown in
[0050] Figure 6 A structural block diagram of a tunnel surrounding rock quality prediction system according to an embodiment of the present application is shown in
[0051] Figure 7 A structural diagram of a computing device suitable for implementing embodiments of the present application is shown. DETAILED DESCRIPTION
[0052] The present application will be further described below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely intended for the purpose of explaining the present application and are not intended to limit the present application. In addition, it should be noted that only parts related to the present application are shown in the accompanying drawings for the purpose of description.
[0053] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
[0054] A tunnel surrounding rock quality prediction method and system according to embodiments of the present application will be described below in conjunction with the accompanying drawings.
[0055] The tunnel boring machine (TBM) tunneling parameter prediction method, system and device according to embodiments of the present application can realize TBM tunneling surrounding rock quality prediction based on multi-source data fusion, solving the technical problems of single data form and inaccurate prediction in related technologies.
[0056] Figure 1 A flowchart of a tunnel surrounding rock quality prediction method according to an embodiment of the present application is shown in Figure 1 A tunnel surrounding rock quality prediction method according to an embodiment of the present application is shown in Figure 2 The tunnel surrounding rock quality prediction method according to an embodiment of the present application includes the following steps:
[0057] S101: Obtain actual multi-source data in the tunnel boring machine tunneling process, wherein the multi-source data at least includes surrounding rock slag image data, tunnel boring machine tunneling data and vibration signal data.
[0058] Specifically, real-time multi-source data in the tunnel boring machine TBM tunneling process is collected, including: tunnel boring machine tunneling data (TBM tunneling data), slag image data and tunnel boring machine vibration signal data (TBM vibration signal), wherein the TBM tunneling data includes but is not limited to pushing speed, cutter head rotating speed, pushing force, torque, penetration, etc. The slag image data can be collected by a high-definition camera installed on the conveyor belt. The TBM vibration signal can be collected by an acceleration sensor arranged at the cutter head, main beam and other key positions.
[0059] S102: Extract the rock slag image data features, tunneling data features and vibration signal data features of the rock slag image data respectively.
[0060] S103: Perform feature fusion on the rock slag image data features, tunneling data features and vibration signal data features of the rock slag image data, and obtain the final prediction result of the surrounding rock quality according to the fused features, or respectively obtain the first prediction result of the surrounding rock quality according to the rock slag image data features, the second prediction result of the surrounding rock quality according to the tunneling data features, and the third prediction result of the surrounding rock quality according to the vibration signal data features, and decide the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
[0061] In an embodiment of the present application, the feature fusion of the rock slag image data features, tunneling data features and vibration signal data features of the rock slag image data, and the final prediction result of the surrounding rock quality according to the fused features, include: performing feature splicing fusion on the rock slag image data features, tunneling data features and vibration signal data features of the rock slag image data to obtain fused features; inputting the fused features into a pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, wherein the first prediction model is pre-trained with a plurality of multi-source data of the tunnel boring machine tunneling process as training samples and the surrounding rock quality evaluation result as a label.
[0062] In combination with Figure 3 , Figure 4 and Figure 5 , before the fused features are input into the pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, it further includes: a plurality of multi-source data of the tunnel boring machine tunneling process; optimizing each type of multi-source data in the plurality of multi-source data; extracting the features of each type of multi-source data after optimization; performing feature splicing fusion on the features of each type of multi-source data to obtain fused features; training the initial first prediction model according to the fused features to obtain the pre-trained first prediction model.
[0063] Wherein, the optimization processing of each type of multi-source data in the plurality of multi-source data includes: removing outliers in each type of multi-source data; after removing outliers in each type of multi-source data, performing alignment processing on each type of multi-source data.
[0064] After the optimization processing, the features of each type of multi-source data are extracted, including: dividing the tunneling cycle from the tunneling data, and extracting the tunneling data features according to the divided tunneling cycle, wherein the tunneling data features include the mean, variance and range of the tunneling data; extracting the rock residue image data features of the surrounding rock from the rock residue image data of the surrounding rock, the rock residue image data features of the surrounding rock including the size and shape features of each rock block; and extracting the vibration signal data features from the vibration signal data, wherein the vibration signal data features include time domain features and frequency domain features.
[0065] Specifically, the multi-source data collected in the TBM tunneling process is collected, including: TBM tunneling data: push speed, cutter head speed, thrust, torque, penetration, and other parameters; rock residue image: rock residue image collected by a high-definition camera installed on the conveyor belt, extracting rock block size, shape, and other information; TBM vibration signal: vibration time domain and frequency domain signals collected by acceleration sensors arranged at key positions such as the cutter head and the main beam.
[0066] Then, the collected multi-source data is preprocessed, including: (1) data outlier rejection. That is: TBM tunneling parameters adopt a dynamic threshold method: calculate the parameter mean ± 3σ (σ is the standard deviation) with every 10 minutes as the window, and the data outside the range is considered abnormal. Rock residue image: if there is no rock residue image in the image, it is rejected; through similarity comparison of adjacent images, if the similarity is higher than 85%, the repeated image is rejected. Vibration signal: use Hampel filter (window length 50, threshold 3) to eliminate pulse noise. (2) Data alignment processing: on the one hand, select three types of data as tunneling data, rock residue image data, and TBM vibration signal data recorded at the same time during TBM tunneling. The processing method is to correspond according to the time when the tunneling data is recorded, and the three types of data in the same time period are segmented first. On the other hand, the data range is aligned, because the sampling frequencies of the sensors are inconsistent, the number of data recorded by time is inconsistent, and the processing method is to align according to the lowest sampling frequency, and in the case of high-frequency sampling, the average value in the low-frequency sampling time interval is calculated.
[0067] Finally, the feature indicators capable of representing the quality of surrounding rock are extracted from each type of data, as follows: TBM tunneling data: push speed, cutter head speed, thrust, torque, penetration, and other parameters; dividing the tunneling cycle from the TBM tunneling data, and extracting statistical features (mean, variance, extreme value, etc.), as follows:
[0068] First, the preprocessed TBM tunneling data is extracted, and the data value of the cutter thrust is picked up according to time, if the cutter thrust changes from 0 to a non-0 value, the time is the starting point of tunneling, and if the cutter thrust changes from a non-0 value to 0, the time is the end point of tunneling.
[0069] Secondly, based on the data table of the tunneling cycle section, the average value of the thrust of the stable section is calculated (only the thrust data in the range of 25% to 75% is considered for average value calculation), and the time is traversed from front to back, the first thrust greater than the average thrust is recorded as the stable section start time, and the time is traversed from back to front, the first thrust greater than the average thrust is recorded as the stable section end time.
[0070] Finally, according to the four times picked up, a complete tunneling cycle is divided into the rising section, the stable section and the falling section, and the tunneling data mean, variance and range of the stable section are calculated.
[0071] Rock image: each rock image contains a plurality of rock blocks, and the size, shape and other information of the rock blocks are counted;
[0072] The size parameters of the rock block include:
[0073] Length (mm) Length of the longest axis of the rock block
[0074] Width (mm) Maximum width of the rock block perpendicular to the length axis
[0075] Area (mm2) Area of the closed contour of the rock block
[0076] Equivalent diameter (mm) Diameter of a circle equal to the area of the rock block
[0077] Perimeter (mm) Perimeter of the closed contour of the rock block
[0078] The shape parameters of the rock block include:
[0079] Slenderness Length-width ratio
[0080] Circularity Area of the rock block to the area of a circle with the same perimeter
[0081] The TBM vibration signal includes: vibration time domain and frequency domain signal.
[0082] According to the pretreated triaxial time domain data, the triaxial peak-to-peak value (PP), root mean square value (RMS), kurtosis coefficient (K), skewness coefficient (S) and margin coefficient (C) are extracted by batch processing; in the frequency domain, the triaxial vibration signal is converted into a frequency domain signal by using Fourier transform, and then the spectral center frequency fc, low frequency energy ratio Rlow, wavelet packet energy entropy H and CMV index of the signal are extracted by batch processing.
[0083] The processed data is stored in a unified format to form a structured database.
[0084] In an embodiment of the present application, the initial first prediction model is trained according to the fusion features to obtain the pre-trained first prediction model, including: inputting the fusion features into the initial first prediction model, and training the initial first prediction model according to the loss of the output of the initial first prediction model and the label until the loss of the output of the initial first prediction model and the label meets the set requirement, to obtain the pre-trained first prediction model.
[0085] Specifically, the preprocessed data is divided into a training set and a test set (ratio 8:2), and a multi-source data fusion algorithm is used for model training, including: feature dimension fusion, that is, the feature vectors of different data sources are fused in the input layer; feature extraction is performed on three types of data (TBM tunneling data, rock slag image, and vibration signal); the extracted feature vectors are spliced and fused in the input layer; the fused feature vectors are input into a unified model for training; the specific fusion method is:
[0086] Input feature = [TBM tunneling data feature] + [rock slag image data feature] + [TBM vibration signal data feature]. Then, the model is trained based on the fused features. After training, the actual multi-source data obtained during the tunneling process of the tunnel boring machine can be output through the model to obtain the final prediction result of the surrounding rock quality.
[0087] In another embodiment, the first prediction result of the surrounding rock quality can also be obtained according to the rock slag image data feature, the second prediction result of the surrounding rock quality can be obtained according to the tunneling data feature, and the third prediction result of the surrounding rock quality can be obtained according to the vibration signal data feature, and the final prediction result of the surrounding rock quality is determined according to the first prediction result, the second prediction result, and the third prediction result, specifically including: inputting the rock slag image data feature into a pre-trained second prediction model to obtain the first prediction result of the surrounding rock quality; inputting the tunneling data feature into a pre-trained third prediction model to obtain the second prediction result of the surrounding rock quality; inputting the vibration signal data feature into a pre-trained fourth prediction model to obtain the third prediction result of the surrounding rock quality; and decision fusion is performed on the first prediction result, the second prediction result, and the third prediction result to determine the final prediction result of the surrounding rock quality.
[0088] In this example, the decision fusion of the first prediction result, the second prediction result, and the third prediction result is performed to determine the final prediction result of the surrounding rock quality, including: voting on the first prediction result, the second prediction result, and the third prediction result respectively; and the prediction result with the most votes is taken as the final prediction result of the surrounding rock quality.
[0089] Specifically, the prediction results of different data sources are voted.
[0090] Independent models are trained with three types of data respectively, and the prediction result with more votes is taken as the final result. For example, the surrounding rock grade obtained by the excavation data is IV, the surrounding rock grade obtained by the rock slag image is III, and the surrounding rock grade obtained by the vibration signal is IV, and the final output result is IV.
[0091] It should be noted that the prediction model of the present application outputs the surrounding rock grade, and the surrounding rock grade can be directly evaluated after the TBM excavation, which is used as the comparison of the prediction result of the training set.
[0092] The trained model is used to predict the real-time collected data, and the surrounding rock quality evaluation result is output, specifically the surrounding rock grade. According to the classification table of surrounding rock shown in Table 1, the surrounding rock grade is divided into 5 types in this example.
[0093] Table 1
[0094]
[0095] After the final prediction result of the surrounding rock quality, it further includes: adjusting the construction decision of the tunnel boring machine according to the final prediction result of the surrounding rock quality.
[0096] According to the tunnel surrounding rock quality prediction method of the embodiment of the present application, after the actual multi-source data in the tunnel boring machine tunneling process is obtained, the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features are extracted, finally, the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features are fused, and the final prediction result of the surrounding rock quality is obtained according to the fused features, or, the first prediction result of the surrounding rock quality is obtained according to the rock slag image data features, the second prediction result of the surrounding rock quality is obtained according to the tunneling data features, and the third prediction result of the surrounding rock quality is obtained according to the vibration signal data features, and the final prediction result of the surrounding rock quality is decided according to the first prediction result, the second prediction result and the third prediction result. By reasonably fusing the multi-source information such as the TBM tunneling data, the rock slag image and the vibration signal, more comprehensive and accurate evaluation of the surrounding rock quality can be realized, and the following advantages are obtained: high comprehensive evaluation, fusion of three complementary data sources of the TBM tunneling data, the rock slag image and the vibration signal, overcoming the limitations of single data evaluation; high accuracy, by means of the feature fusion and decision fusion strategies, the associated information between the data is fully mined, and the evaluation accuracy is improved; strong real-time, based on the real-time collected data for rapid analysis, meeting the real-time decision-making needs of the TBM construction, strong self-adaptation, the intelligent algorithm can adapt to different geological conditions, and has strong generalization ability; good engineering practicability, completely based on the TBM construction process data, without additional exploration, and the construction cost is reduced.
[0097] Figure 6 is a structural block diagram of a tunnel surrounding rock quality prediction system according to an embodiment of the present application, as Figure 6 shown, the tunnel surrounding rock quality prediction system according to an embodiment of the present application comprises: an acquisition module 610, an extraction module 620 and a prediction module 630, wherein:
[0098] The acquisition module 610 is used for obtaining actual multi-source data in the tunnel boring machine tunneling process, wherein the multi-source data at least comprises rock slag image data of surrounding rock, tunneling data of the tunnel boring machine and vibration signal data;
[0099] The extraction module 620 is used for extracting the rock slag image data features of the rock slag image data, the tunneling data features and the vibration signal data features respectively;
[0100] The prediction module 630 is configured to perform feature fusion on the rock slag image data features, the tunneling data features and the vibration signal data features of the rock slag image data, and obtain a final prediction result of the surrounding rock quality according to the fused features, or obtain a first prediction result of the surrounding rock quality according to the rock slag image data features, a second prediction result of the surrounding rock quality according to the tunneling data features and a third prediction result of the surrounding rock quality according to the vibration signal data features, and determine the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
[0101] According to the tunnel surrounding rock quality prediction system provided in the embodiments of the present application, after obtaining the actual multi-source data in the tunnel boring machine tunneling process, the rock slag image data features, the tunneling data features and the vibration signal data features are extracted, and finally, the rock slag image data features, the tunneling data features and the vibration signal data features are fused, and the final prediction result of the surrounding rock quality is obtained according to the fused features, or the first prediction result of the surrounding rock quality is obtained according to the rock slag image data features, the second prediction result of the surrounding rock quality is obtained according to the tunneling data features, and the third prediction result of the surrounding rock quality is obtained according to the vibration signal data features, and the final prediction result of the surrounding rock quality is determined according to the first prediction result, the second prediction result and the third prediction result. By reasonably fusing the multi-source information such as the TBM tunneling data, the rock slag image and the vibration signal, more comprehensive and accurate evaluation of the surrounding rock quality can be realized, and the following advantages are achieved: high comprehensive evaluation, fusion of three complementary data sources of the TBM tunneling data, the rock slag image and the vibration signal, overcoming the limitations of single data evaluation; high accuracy, by means of feature fusion and decision fusion strategies, the associated information between the data is fully mined, and the evaluation accuracy is improved; strong real-time performance, based on real-time data acquisition for rapid analysis, meeting the real-time decision-making needs of TBM construction, strong self-adaptation, intelligent algorithm can adapt to different geological conditions, and has strong generalization ability; good engineering practicability, completely based on TBM construction process data, without additional exploration, and the construction cost is reduced.
[0102] It should be noted that the specific implementation manner of the tunnel surrounding rock quality prediction system in the embodiments of the present application is similar to that of the tunnel surrounding rock quality prediction method in the embodiments of the present application, and specific details are described in the method part, which will not be repeated here.
[0103] Reference will be made to the following description of the embodiments of the present application Figure 7 , Figure 7 A structural schematic diagram of a computing device suitable for implementing the embodiments of the present application is shown.
[0104] As Figure 7As shown, the computer system includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 1002 or programs loaded from storage section 1008 into random access memory (RAM) 1003. RAM 1003 also stores various programs and data required for the system's operating instructions. CPU 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. Input / output (I / O) interface 1005 is also connected to bus 1004.
[0105] 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.
[0106] 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.
[0107] It should be noted that the computer-readable medium shown in the application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, the 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, device or component. In this application, the computer-readable signal medium can include a data signal carried in a baseband or as a carrier wave part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take many forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or component. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
[0108] The flowcharts and block diagrams in the drawings illustrate the possible implementation architectures, functions and operation instructions of the systems, methods and computer program products according to various embodiments of the application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment or a part of code, which contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can also occur in different order from that noted in the drawings. For example, two connected blocks can actually be executed substantially in parallel, and sometimes in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operation instructions, or can be implemented by a combination of dedicated hardware and computer instructions.
[0109] The units or modules described in the embodiments of the present application can be implemented in the form of software or in the form of hardware. The described units or modules can also be arranged in a processor. In some cases, the names of the units or modules do not constitute a limitation on the units or modules themselves.
[0110] As another aspect, the present application also provides a computer readable storage medium, which can be included in the computing device described in the above embodiments, or can exist separately without being assembled into the computing device. The computer readable storage medium stores one or more programs, when the programs are used by one or more processors to execute the tunnel surrounding rock quality prediction method described in the present application. That is: obtaining actual multi-source data in the tunnel boring machine tunneling process, wherein the multi-source data at least includes rock slag image data of surrounding rock, tunneling data and vibration signal data of the tunnel boring machine;
[0111] Respectively extracting rock slag image data features of the rock slag image data, tunneling data features and vibration signal data features;
[0112] Performing feature fusion on the rock slag image data features, the tunneling data features and the vibration signal data features of the rock slag image data, and obtaining the final prediction result of the surrounding rock quality according to the fused features, or respectively obtaining the first prediction result of the surrounding rock quality according to the rock slag image data features, the second prediction result of the surrounding rock quality according to the tunneling data features, and the third prediction result of the surrounding rock quality according to the vibration signal data features, and deciding the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
[0113] The above description is only the preferred embodiments of the present application and the explanation of the applied technical principles. Those skilled in the art should understand that the disclosed range in the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and also covers other technical solutions formed by any combinations of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features disclosed in the present application (but not limited to) having similar functions to form technical solutions.
Claims
1. A tunnel surrounding rock quality prediction method, characterized in that, The method comprises the following steps: obtaining actual multi-source data in a tunnel boring machine tunneling process, wherein the multi-source data at least includes rock residue image data of surrounding rock, tunneling data of the tunnel boring machine and vibration signal data; extracting rock residue image data features of the rock residue image data, tunneling data features and vibration signal data features respectively; performing feature fusion on the rock residue image data features, the tunneling data features and the vibration signal data features, and obtaining a final prediction result of the surrounding rock quality according to the fused features, or obtaining a first prediction result of the surrounding rock quality according to the rock residue image data features, a second prediction result of the surrounding rock quality according to the tunneling data features and a third prediction result of the surrounding rock quality according to the vibration signal data features, and deciding the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
2. The tunnel surrounding rock quality prediction method according to claim 1, characterized by, The feature fusion on the rock residue image data features, the tunneling data features and the vibration signal data features, and the obtaining of the final prediction result of the surrounding rock quality according to the fused features comprise: performing feature splicing fusion on the rock residue image data features, the tunneling data features and the vibration signal data features to obtain fused features; inputting the fused features into a pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, wherein the first prediction model is pre-trained with a plurality of multi-source data of a tunnel boring machine tunneling process as training samples and with a surrounding rock quality evaluation result as a label.
3. The tunnel surrounding rock quality prediction method according to claim 2, characterized by, Before the inputting of the fused features into the pre-trained first prediction model to obtain the final prediction result of the surrounding rock quality, the method further comprises the following steps: a plurality of multi-source data of a tunnel boring machine tunneling process; optimizing each type of multi-source data in the plurality of multi-source data; extracting features of each type of multi-source data after the optimization; performing feature splicing fusion on the features of each type of multi-source data to obtain fused features; training an initial first prediction model according to the fused features to obtain the pre-trained first prediction model.
4. The tunnel surrounding rock quality prediction method according to claim 3, characterized by, The optimization of each type of multi-source data in the plurality of multi-source data comprises: eliminating outliers in each type of multi-source data; aligning each type of multi-source data after the elimination of outliers in each type of multi-source data.
5. The tunnel surrounding rock quality prediction method according to claim 3, characterized by, The extraction of features of each type of multi-source data after the optimization comprises: dividing tunneling cycles from the tunneling data, and extracting tunneling data features according to the divided tunneling cycles, wherein the tunneling data features include mean, variance and range of the tunneling data; extracting rock residue image data features of the surrounding rock from the rock residue image data of the surrounding rock, wherein the rock residue image data features of the surrounding rock include size and shape features of each rock block; extracting vibration signal data features from the vibration signal data, wherein the vibration signal data features include time domain features and frequency domain features.
6. The tunnel surrounding rock quality prediction method according to any one of claims 3-5, characterized in that, The training of the initial first prediction model according to the fused features to obtain the pre-trained first prediction model comprises: The fusion features are input into the initial first prediction model, and the initial first prediction model is trained according to the output of the initial first prediction model and the loss before the label until the output of the initial first prediction model and the loss before the label meet the set requirement, so as to obtain a pre-trained first prediction model.
7. The tunnel surrounding rock quality prediction method according to claim 1, characterized by, The first prediction result of the surrounding rock quality according to the rock residue image data features, the second prediction result of the surrounding rock quality according to the tunneling data features, and the third prediction result of the surrounding rock quality according to the vibration signal data features are obtained, and the final prediction result of the surrounding rock quality is decided according to the first prediction result, the second prediction result and the third prediction result, which comprises: The rock residue image data features are input into a pre-trained second prediction model to obtain the first prediction result of the surrounding rock quality; The tunneling data features are input into a pre-trained third prediction model to obtain the second prediction result of the surrounding rock quality; The vibration signal data features are input into a pre-trained fourth prediction model to obtain the third prediction result of the surrounding rock quality. The first prediction result, the second prediction result and the third prediction result are decision-fused to determine the final prediction result of the surrounding rock quality.
8. The tunnel surrounding rock quality prediction method according to claim 7, characterized by, The first prediction result, the second prediction result and the third prediction result are decision-fused to determine the final prediction result of the surrounding rock quality, which comprises: The first prediction result, the second prediction result and the third prediction result are voted respectively; The prediction result with the most votes is taken as the final prediction result of the surrounding rock quality.
9. The tunnel surrounding rock quality prediction method according to claim 1, characterized by, After the final prediction result of the surrounding rock quality, there further comprises: According to the final prediction result of the surrounding rock quality, the construction decision of the tunnel boring machine is adjusted. 10.A system for predicting quality of surrounding rock of a tunnel, the system comprising: It comprises: An acquisition module is configured to obtain actual multi-source data in a tunnel boring process of a tunnel boring machine, wherein the multi-source data at least comprises rock residue image data of surrounding rock, tunneling data of the tunnel boring machine and vibration signal data; An extraction module is configured to extract rock residue image data features, tunneling data features and vibration signal data features of the rock residue image data respectively; A prediction module is configured to perform feature fusion on the rock residue image data features, the tunneling data features and the vibration signal data features of the rock residue image data, and obtain a final prediction result of the surrounding rock quality according to the fusion features, or obtain a first prediction result of the surrounding rock quality according to the rock residue image data features, a second prediction result of the surrounding rock quality according to the tunneling data features and a third prediction result of the surrounding rock quality according to the vibration signal data features, and decide the final prediction result of the surrounding rock quality according to the first prediction result, the second prediction result and the third prediction result.
Citation Information
Cited By
Control method and device of excavation system, equipment and medium
CN121918385A