Real-time detection system and deduction method for shield muck fine-grained soil content based on multi-parameter fusion

Abstract

This invention relates to the field of tunnel boring machine (TBM) construction technology, specifically to a real-time detection system and deduction method for fine-grained soil content in TBM excavated soil based on multi-parameter fusion. The system includes a multi-parameter sensor unit, a data acquisition and transmission unit, a data processing and analysis unit, and a display unit. In this invention, the system rapidly detects the fine-grained soil content in the slurry separated from the TBM excavated soil using an intelligent density meter after sedimentation. A correction algorithm based on the data difference law obtained from indoor experiments is introduced to correct the fine-grained soil content data obtained from the intelligent density meter. Using a multi-parameter fusion algorithm, the system comprehensively analyzes various processed TBM excavated soil data, achieving rapid detection and analysis of changes in the fine-grained soil component in the TBM excavated soil. This system solves the problems of long processing time, complex operation, and poor adaptability of traditional density meter methods. The detection system enables real-time monitoring of fine-grained soil content, providing a basis for the resource utilization of excavated soil.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine (TBM) construction technology, specifically to a real-time detection system and deduction method for the fine-grained soil content of TBM excavation soil based on multi-parameter fusion. Background Technology

[0002] During shield tunnel construction, the properties of the excavated soil have a crucial impact on construction safety, efficiency, and project quality. Among these, the fine-grained soil content is a key parameter, directly affecting the soil's fluidity, stability, and the degree of wear on the shield equipment. Accurately determining the fine-grained soil content in the shield excavated soil helps construction personnel adjust construction parameters in a timely manner, such as excavated soil improvement schemes, shield advancement speed, and control of excavated soil volume, thereby ensuring smooth construction and reducing construction risks and costs.

[0003] Existing sensor-based detection methods often only detect a single parameter, making it difficult to comprehensively and accurately reflect the true content of fine-grained soil in the excavated soil. Due to the complex environment of shield tunneling and the fact that the properties of the excavated soil are affected by a variety of factors, the detection of a single parameter is easily interfered with, resulting in inaccurate detection results.

[0004] Furthermore, traditional methods for detecting fine soil content based on densitometers are time-consuming, require highly skilled operators, and cannot adapt to rapidly changing geological conditions. However, during shield tunneling, changes in the composition of fine soil in the excavated soil not only affect the construction measures for excavated soil improvement during tunneling but also the process design for subsequent excavated soil reuse. Therefore, this application proposes a real-time detection system and deduction method for fine soil content in shield tunnel excavated soil based on multi-parameter fusion. Summary of the Invention

[0005] The purpose of this invention is to provide a real-time detection system and deduction method for the fine soil content of shield tunnel slag based on multi-parameter fusion, so as to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The real-time detection system for fine-grained soil content in shield tunnel slag based on multi-parameter fusion includes a multi-parameter sensor unit, a data acquisition and transmission unit, a data processing and analysis unit, and a display unit.

[0008] The multi-parameter sensor unit measures the particle size distribution of the slag particles in real time using a laser particle size analyzer, detects the moisture content of the slag in real time using a moisture content sensor, detects and determines the density of the slag using a density sensor, separates the slag to obtain mud and water using a mud-water separation device, and detects the density of the settled mud and water using an intelligent density meter.

[0009] The data acquisition and transmission unit is used to acquire electrical or digital signals output by the multi-parameter sensor unit, process them, and then transmit the shield tunneling excavation data to the data processing and analysis unit in real time.

[0010] The data processing and analysis unit filters and denoises the shield tunneling excavation data, introduces a correction algorithm based on the data difference law obtained from indoor tests, corrects the fine soil content data detected by the intelligent density meter, and uses a multi-parameter fusion algorithm to comprehensively analyze the shield tunneling excavation data after various processing. Combining the excavation weight and coarse particle gradation parameters, the fine soil content in the shield tunneling excavation is calculated.

[0011] The display unit displays the calculation results of the fine soil content of the slag, the real-time measurement data of each sensor, the weight of the slag, the coarse particle size distribution, and related construction parameters in real time. When the detected fine soil content exceeds the preset reasonable range, the alarm will immediately issue an audible and visual alarm signal.

[0012] Furthermore, the multi-parameter sensor unit includes a particle size analysis module, a moisture content detection module, a slag density detection module, a mud-water separation module, and a mud-water density detection module;

[0013] The particle size analysis module uses a laser particle size analyzer to measure the particle size distribution of slag particles in real time, and obtains the particle size ratio of fine soil in the slag.

[0014] The moisture content detection module uses a moisture content sensor to detect the moisture content of the slag in real time;

[0015] The slag density detection module detects and determines the slag density using a density sensor.

[0016] The mud-water separation module separates the slag and soil to obtain mud and water through a mud-water separation device;

[0017] The mud-water density detection module uses an intelligent density meter to detect the density of the settled mud-water. Based on the correlation between mud-water density and fine-grained soil content, it calculates the content of fine-grained soil in the mud-water.

[0018] Furthermore, in the data acquisition and transmission unit, the electrical or digital signals output by the multi-parameter sensor unit are converted into a unified data format, a high-precision A / D converter is used to digitize the analog signals, the multi-sensor data is integrated through a data acquisition card, and the acquired data is transmitted to the data processing and analysis unit in real time using wireless communication technology.

[0019] Furthermore, the data processing and analysis unit includes a preliminary verification module, a data processing module, a data alignment module, a data correction module, a multi-parameter feature extraction module, a multi-parameter fusion algorithm operation module, and a data management module;

[0020] The preliminary verification module receives data from the data acquisition and transmission unit in real time through a preset communication interface protocol. During the reception process, it performs preliminary verification on the integrity and format of the data. For data with problems, it automatically sends a retransmission request to the data acquisition and transmission unit.

[0021] The data processing module employs various filtering algorithms to process different types of shield tunneling excavation data;

[0022] The data alignment module uses time synchronization technology to calibrate and align the timestamps of various shield tunnel excavation data from various sensors, so that all parameters are consistent in the time dimension and form a unified time series dataset.

[0023] The data correction module corrects the fine-grained soil content data obtained by the smart densitometer by introducing a correction algorithm based on the data difference pattern obtained from indoor tests.

[0024] The multi-parameter feature extraction module extracts features from the preprocessed and corrected data, mines the feature information related to the fine soil content of each parameter, calculates the mass of fine soil in a unit volume of slag by combining the slag weight data, and calculates the ratio of coarse particles to fine soil by combining the coarse particle gradation data, forming a multi-dimensional feature vector.

[0025] The multi-parameter fusion algorithm calculation module uses a multi-parameter fusion algorithm to comprehensively analyze the extracted multi-dimensional feature vectors and calculate the content of fine-grained soil in the shield tunnel slag.

[0026] The data management module uses a distributed database architecture to store the processed data, including raw sensor data, preprocessed data, correction results, fusion calculation results, and related feature parameters. It also regularly backs up the data to local hard drives and cloud storage.

[0027] Furthermore, in the data processing module, for continuous data output by the laser particle size analyzer, moisture content sensor, and density sensor, a median filtering algorithm is used to remove instantaneous pulse interference; for data output by the intelligent density meter, a Kalman filtering algorithm is used to predict the expected value of the data by establishing a dynamic model, and to dynamically correct it based on the deviation between the actual measured value and the predicted value.

[0028] Furthermore, the data correction module corrects the fine-grained soil content data obtained by the smart density meter by introducing a correction algorithm based on the data difference pattern obtained from indoor tests, including the following steps:

[0029] A1. During the indoor testing phase, a large number of slag soil samples with different fine-grained soil contents were collected and tested using both the intelligent density meter method and the traditional laboratory sieving method. A database of the differences between the test results of the two methods was established.

[0030] A2. In the correction algorithm, the real-time detection value of the smart densitometer is used as input. By querying the difference database and combining it with a linear regression model or a BP neural network model, the corresponding correction coefficient is calculated to correct the detection value.

[0031] A3. Regularly update the difference database, incorporating the latest field test data and laboratory comparison data, and continuously optimize the parameters of the correction model.

[0032] The derivation method of a real-time detection system for fine-grained soil content in shield tunnel excavation soil based on multi-parameter fusion includes the following steps:

[0033] I. Establishing a historical database: During the tunnel boring machine (TBM) construction process, the system continuously collects and stores multi-parameter detection data, TBM tunneling parameters, and geological condition information at different construction stages to establish a historical database;

[0034] II. Data Feature Extraction: In-depth analysis of historical database data is conducted to extract feature parameters closely related to changes in fine-grained soil content. Statistical analysis methods are used to determine which parameters are strongly correlated with fluctuations in fine-grained soil content under different geological conditions.

[0035] III. Data Feature Selection: Data mining techniques are used to uncover potential correlation features between various parameters. Feature selection algorithms are then used to select the most representative subset of feature parameters that contribute the most to the inference of fine-grained soil content from among the numerous extracted features.

[0036] IV. Constructing the projection model: Based on the selected feature parameters, a fine-grained soil content projection model is constructed using machine learning algorithms or time series analysis methods;

[0037] V. Real-time simulation and result output: During the shield tunneling process, the current multi-parameter detection data, muck weight, coarse particle size distribution, shield tunneling parameters, and geological condition information are input into the validated and optimized simulation model in real time. Based on the input data, the model quickly predicts the changing trend of fine soil content in the shield muck over the next 5-10 rings of tunneling time and outputs the simulation results to the display unit for display.

[0038] Furthermore, step IV includes the following steps:

[0039] B1. Based on the selected feature parameters, a model for extrapolating the content of fine-grained soil is constructed using machine learning algorithms or time series analysis methods.

[0040] B2. During the model building process, a portion of the data in the historical database is used as the training set to optimize and adjust the model parameters;

[0041] B3. Use the remaining data in the historical database as a test set to verify the constructed inference model. Evaluate the model's predictive performance by calculating the error index between the model's prediction results and the actual fine-grained soil content data in the test set.

[0042] B4. If the model error is large, further analyze the reasons, take corresponding optimization measures, and then train and validate the model again until the model achieves satisfactory prediction accuracy and stability.

[0043] The beneficial effects of this invention are:

[0044] 1. In this invention, the system rapidly detects the fine-grained soil content in the slurry separated from the tunnel boring machine (TBM) excavation soil by using an intelligent density meter after sedimentation. A correction algorithm based on the data difference pattern obtained from indoor experiments is introduced to correct the fine-grained soil content data detected by the intelligent density meter. A multi-parameter fusion algorithm is used to comprehensively analyze various processed TBM excavation soil data, combining the excavation soil weight and coarse particle size distribution parameters to achieve rapid detection and analysis of changes in the fine-grained soil component in the TBM excavation soil. This allows for more timely reflection of changes in the properties of the excavation soil and fully utilizes the complementarity between various parameters. Compared with the original method, it more comprehensively and accurately reflects the true content of fine-grained soil, improving the reliability of the detection results. It achieves the fusion of multiple rapid detection methods. This system solves the problems of long processing time, complex operation, and poor adaptability of the traditional density meter method. The detection system enables real-time monitoring of fine-grained soil content, providing a basis for the resource utilization of excavation soil.

[0045] 2. In this invention, the deduction method of the real-time detection system for fine soil content of shield tunneling slag based on multi-parameter fusion is achieved by establishing a historical database, extracting data features, selecting data features, constructing a deduction model, and performing real-time deduction and outputting results. This deduction model integrates various newly added parameters, which can more comprehensively predict the trend of fine soil content changes, enhance the controllability and safety of construction, and promote the improvement of the level of intelligent shield tunneling construction. Attached Figure Description

[0046] Figure 1 This is a system schematic diagram of the real-time detection system for fine-grained soil content in shield tunnel slag based on multi-parameter fusion, as described in this invention.

[0047] Figure 2 This is a flowchart illustrating the derivation method of the real-time detection system for fine-grained soil content in shield tunnel slag based on multi-parameter fusion, as described in this invention.

[0048] In the diagram: 1. Multi-parameter sensor unit; 2. Data acquisition and transmission unit; 3. Data processing and analysis unit; 4. Display unit; 11. Particle size analysis module; 12. Moisture content detection module; 13. Slag density detection module; 14. Mud-water separation module; 15. Mud-water density detection module; 31. Preliminary verification module; 32. Data processing module; 33. Data alignment module; 34. Data correction module; 35. Multi-parameter feature extraction module; 36. Multi-parameter fusion algorithm operation module; 37. Data management module. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely 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 without creative effort are within the scope of protection of the present invention.

[0050] Example 1: Please refer to Figure 1 and Figure 2 This design proposes an implementation method for a real-time detection system for fine-grained soil content in shield tunnel slag based on multi-parameter fusion, comprising a multi-parameter sensor unit 1, a data acquisition and transmission unit 2, a data processing and analysis unit 3, and a display unit 4.

[0051] The multi-parameter sensor unit 1 measures the particle size distribution of the slag particles in real time using a laser particle size analyzer, detects the moisture content of the slag in real time using a moisture content sensor (using a capacitive or microwave sensor), detects and determines the density of the slag using a density sensor, separates the slag to obtain mud and water using a mud-water separation device, and detects the density of the settled mud and water using an intelligent density meter. The multi-parameter sensor unit 1 includes a particle size analysis module 11, a moisture content detection module 12, a slag density detection module 13, a mud-water separation module 14, and a mud-water density detection module 15.

[0052] The particle size analysis module 11 measures the particle size distribution of the slag particles in real time using a laser particle size analyzer to obtain the particle size ratio of fine soil in the slag; the moisture content detection module 12 detects the moisture content of the slag in real time using a moisture content sensor; the slag density detection module 13 detects and determines the density of the slag using a density sensor; the mud-water separation module 14 separates the slag into mud and water using a mud-water separation device; and the mud-water density detection module 15 detects the density of the settled mud and water using an intelligent density meter, and calculates the content of fine soil in the mud and water based on the correspondence between the mud and water density and the content of fine soil.

[0053] The data acquisition and transmission unit 2 is used to acquire and process the electrical or digital signals output by the multi-parameter sensor unit 1. This includes converting the electrical or digital signals output by the multi-parameter sensor unit 1 into a unified data format, using a high-precision A / D converter to digitize the analog signals, integrating the multi-sensor data through a data acquisition card, using wireless communication technology to transmit the acquired data to the data processing and analysis unit 3 in real time, and finally transmitting the shield tunneling excavation data to the data processing and analysis unit 3 in real time.

[0054] The data processing and analysis unit 3 filters and denoises the shield tunneling slag data, introduces a correction algorithm based on the data difference law obtained from indoor tests, corrects the fine soil content data detected by the intelligent density meter, and uses a multi-parameter fusion algorithm to comprehensively analyze the shield tunneling slag data after various processing. Combining the slag weight and coarse particle gradation parameters, it calculates the fine soil content in the shield tunneling slag. The data processing and analysis unit 3 includes a preliminary verification module 31, a data processing module 32, a data alignment module 33, a data correction module 34, a multi-parameter feature extraction module 35, a multi-parameter fusion algorithm operation module 36, and a data management module 37.

[0055] The preliminary verification module 31 receives data from the data acquisition and transmission unit 2 in real time through a preset communication interface protocol. During the receiving process, it performs preliminary verification on the integrity and format of the data, and checks whether there are problems such as packet loss, missing fields or format errors. For data with problems, it automatically sends a retransmission request to the data acquisition and transmission unit 2.

[0056] The data processing module 32 addresses data fluctuations and outliers caused by factors such as electromagnetic interference and mechanical vibration in the shield tunneling construction environment. It employs various filtering algorithms to process different types of shield tunneling excavation data. For continuous data output from the laser particle size analyzer, moisture content sensor, and density sensor, a median filtering algorithm is used to remove transient pulse interference, effectively eliminating isolated outliers while preserving the data's trend characteristics. For data output from the intelligent density meter, which has a high detection frequency and is easily affected by the slurry flow state, a Kalman filtering algorithm is used. This algorithm establishes a dynamic model to predict the expected value of the data and dynamically corrects it based on the deviation between the actual measured value and the predicted value, thereby smoothing data fluctuations and improving data stability.

[0057] The data alignment module 33 software adopts high-precision time synchronization technology (due to the different detection principles and response speeds of different sensors, the timestamps of their output data may differ. To ensure the accuracy of multi-parameter fusion analysis, the system uses the high-precision clock inside the system as a reference to calibrate and align the timestamps of various shield tunneling excavation data from each sensor. Linear interpolation or spline interpolation algorithms are used to complete the data, so that all parameters are consistent in the time dimension, forming a unified time series dataset.

[0058] The data correction module 34 corrects the fine-grained soil content data obtained by the smart density meter by introducing a correction algorithm based on the data difference pattern obtained from indoor tests. The correction includes the following steps:

[0059] A1. During the indoor testing phase, a large number of slag soil samples with different fine-grained soil contents were collected and tested using both the intelligent density meter method and the traditional laboratory sieving method (as the standard method). A database of the differences between the test results of the two methods was established.

[0060] A2. In the correction algorithm, the real-time detection value of the intelligent density meter is used as input. By querying the difference database and combining it with a linear regression model or a BP neural network model, the corresponding correction coefficient is calculated to correct the detection value, so as to eliminate the systematic error between this rapid detection method and the traditional detection method.

[0061] A3. Regularly update the difference database, incorporate the latest on-site testing data and laboratory comparison data, and continuously optimize the parameters of the correction model to ensure that the correction accuracy continues to improve with the construction process;

[0062] The multi-parameter feature extraction module 35 extracts features from the preprocessed and corrected data, mining the feature information related to the fine-grained soil content of each parameter (for the particle size distribution data output by the laser particle size analyzer, extracting feature parameters such as the volume ratio of fine-grained soil particles, standard deviation of particle size distribution, and characteristic particle size; for the moisture content and density data, extracting statistical features such as mean, variance, and rate of change; for the fine-grained soil content data corrected by the intelligent density meter, extracting features such as instantaneous value, moving average, and cumulative change), calculates the mass of fine-grained soil per unit volume of slag and soil weight data, and calculates the ratio of coarse particles to fine-grained soil by combining coarse particle gradation data, forming a multi-dimensional feature vector;

[0063] The multi-parameter fusion algorithm operation module 36 uses a multi-parameter fusion algorithm to comprehensively analyze the extracted multi-dimensional feature vectors and calculate the content of fine-grained soil in the shield tunnel slag.

[0064] The data management module 37 uses a distributed database architecture to store the processed data, including raw sensor data, preprocessed data, correction results, fusion calculation results, and related feature parameters. It also regularly backs up the data to local hard drives and cloud storage.

[0065] Display unit 4 displays in real time the calculation results of the fine soil content of the slag, the real-time measurement data of each sensor, the weight of the slag, the coarse particle gradation, and related construction parameters. When the detected fine soil content exceeds the preset reasonable range, the alarm will immediately issue an audible and visual alarm signal to remind the construction personnel to take corresponding measures in time, such as adjusting the amount of slag improvement additives and optimizing the shield tunneling parameters. The alarm threshold can be flexibly set by the construction personnel in the system according to different engineering geological conditions and construction requirements.

[0066] In this embodiment, the system rapidly detects the fine-grained soil content in the slurry separated from the tunnel boring machine (TBM) using an intelligent density meter after sedimentation. A correction algorithm based on data difference patterns obtained from indoor experiments is introduced to correct the fine-grained soil content data obtained from the intelligent density meter. A multi-parameter fusion algorithm is used to comprehensively analyze various processed TBM slag data, combining slag weight and coarse particle size distribution parameters to quickly obtain fine-grained soil content data. After fusion with other parameters, the real-time performance and speed of the detection are further improved, allowing for more timely reflection of changes in slag properties. Simultaneously, the system fully utilizes the complementarity between various parameters, providing a more comprehensive and accurate reflection of the true fine-grained soil content compared to traditional methods, improving the reliability of the detection results. It achieves the integration of multiple rapid detection methods, reducing reliance on traditional laboratory testing and lowering manpower and time costs. This system solves the problems of long processing time, complex operation, and poor adaptability associated with traditional density meter methods. The real-time monitoring of fine-grained soil content through the detection system provides a basis for the resource utilization of TBM slag.

[0067] Example 2: Please refer to Figure 1 and Figure 2 This design proposes an implementation method based on a derivation method for a real-time detection system of fine-grained soil content in shield tunnel slag, which includes the following steps:

[0068] I. Establishing a historical database: During the shield tunneling process, the system continuously collects and stores multi-parameter detection data (including particle size distribution, moisture content, density, fine soil content data detected and corrected by intelligent density meter, slag weight, coarse particle gradation and corresponding fine soil content calculation results), shield tunneling parameters (such as propulsion speed, cutterhead speed, torque, etc.) and geological condition information (such as stratum type, rock and soil physical and mechanical parameters, etc.) at different construction stages, establishing a historical database to provide rich samples for subsequent extrapolation and analysis;

[0069] II. Data Feature Extraction: In-depth analysis of historical database data is conducted to extract feature parameters closely related to changes in fine-grained soil content. In addition to the features mentioned above, these include mud-water related parameters (such as mud-water density, fine-grained soil content in mud-water, etc.), slag weight characteristics, coarse particle gradation characteristics, etc. Statistical analysis methods are used to determine which parameter changes are strongly correlated with fluctuations in fine-grained soil content under different geological conditions.

[0070] III. Data Feature Selection: Data mining techniques are used to uncover potential correlation features between various parameters. Feature selection algorithms are then used to select the most representative subset of feature parameters that contribute the most to the inference of fine-grained soil content from among the numerous extracted features, in order to reduce data dimensionality and improve the computational efficiency and accuracy of the inference model.

[0071] IV. Constructing the projection model: Based on the selected feature parameters, a fine-grained soil content projection model is constructed using machine learning algorithms or time series analysis methods, including the following steps:

[0072] B1. Based on the selected feature parameters, a fine-grained soil content prediction model is constructed using machine learning algorithms or time series analysis methods. For example, Support Vector Regression (SVR), Long Short-Term Memory (LSTM) network models, or Autoregressive Integral Moving Average (ARIMA) models can be used. SVR algorithms predict fine-grained soil content by finding an optimal hyperplane; LSTM models can effectively handle long-term dependencies in time series data and capture the changing trend of fine-grained soil content with time and construction progress; and ARIMA models are suitable for predicting time series data that are stationary or stationary after differencing.

[0073] B2. In the process of building the model, a portion of the data in the historical database is used as the training set to optimize and adjust the parameters of the model so that it can accurately fit the complex relationship between the fine-grained soil content and various characteristic parameters.

[0074] B3. Use the remaining data in the historical database as a test set to verify the constructed inference model. Evaluate the model's predictive performance by calculating the error index between the model's prediction results and the actual fine-grained soil content data in the test set.

[0075] B4. If the model error is large, further analyze the reasons, take corresponding optimization measures, and then train and validate the model again. Calculate the error indices between the model's prediction results and the actual fine-grained soil content data in the test set, such as root mean square error (RMSE) and mean absolute error (MAE), to evaluate the model's predictive performance. If the model error is large and does not meet the actual engineering requirements, further analyze the reasons, which may be due to unreasonable feature selection, improper model parameter settings, or insufficient training data. For these problems, take corresponding optimization measures, such as reselecting features, adjusting model parameters, or increasing the amount of training data. Then train and validate the model again until the model achieves satisfactory prediction accuracy and stability.

[0076] V. Real-time simulation and result output: During the shield tunneling process, the current multi-parameter detection data, slag weight, coarse particle size distribution, shield tunneling parameters, and geological condition information are input into the validated and optimized simulation model in real time. Based on the input data, the model quickly predicts the changing trend of fine soil content in the shield slag during the next 5-10 ring tunneling period and outputs the simulation results to display unit 4 for display.

[0077] In this embodiment, the deduction method of the real-time detection system for fine soil content of shield tunneling excavation soil based on multi-parameter fusion is carried out by establishing a historical database, extracting data features, selecting data features, constructing a deduction model, and performing real-time deduction and outputting results. The deduction model integrates various newly added parameters, which can more comprehensively predict the trend of fine soil content change, enhance the controllability and safety of construction, and promote the improvement of the level of intelligent shield tunneling construction.

[0078] In this invention, the real-time detection system for fine-grained soil content in shield tunnel excavation soil mainly consists of a multi-parameter sensor unit 1, a data acquisition and transmission unit 2, a data processing and analysis unit 3, and a display unit 4. The multi-parameter sensor unit 1 uses multiple sensors to measure the particle size distribution of the excavation soil, detect its moisture content, detect and determine its density, and detect the density of the settled mud-water in real time. The data acquisition and transmission unit 2 transmits the shield tunneling excavation soil data to the data processing and analysis unit 3 in real time. The data processing and analysis unit 3 filters and denoises the shield tunneling excavation soil data and introduces a correction algorithm based on the data difference law obtained from indoor experiments to correct the fine-grained soil content data detected by the intelligent density meter. Using a multi-parameter fusion algorithm, the system comprehensively analyzes various processed shield tunneling excavation data, combining excavation weight and coarse particle size distribution parameters to calculate the fine soil content in the excavation. Finally, display unit 4 displays the calculation results of the fine soil content, real-time measurement data from each sensor, excavation weight, coarse particle size distribution, and related construction parameters in real time. When the detected fine soil content exceeds the preset reasonable range, the alarm immediately issues an audible and visual alarm signal to remind construction personnel to take appropriate measures in a timely manner, such as adjusting the amount of excavation soil improvement additives and optimizing shield tunneling parameters. The alarm threshold can be flexibly set by construction personnel in the system according to different engineering geological conditions and construction requirements.

[0079] This system rapidly detects the fine-grained soil content in the slurry separated from the tunnel boring machine (TBM) excavation soil by using an intelligent density meter after sedimentation. It incorporates a correction algorithm based on data difference patterns obtained from indoor experiments to correct the fine-grained soil content data detected by the intelligent density meter. Furthermore, a multi-parameter fusion algorithm is used to comprehensively analyze various processed TBM excavation soil data, enabling rapid detection and analysis of changes in the fine-grained soil component within the excavation soil. This system solves the problems of time-consuming, complex, and poorly adaptable traditional density meter methods. By enabling real-time monitoring of fine-grained soil content through the detection system, it provides a basis for the resource utilization of TBM excavation soil.

[0080] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A shield muck fine-grained soil content real-time detection system based on multi-parameter fusion, characterized in that, It comprises a multi-parameter sensor unit (1), a data acquisition and transmission unit (2), a data processing and analysis unit (3) and a display unit (4): The multi-parameter sensor unit (1) measures the particle size distribution of the slag soil particles in real time through a laser particle size analyzer, detects the moisture content of the slag soil in real time through a moisture content sensor, detects and determines the density of the slag soil through a density sensor, separates the slag soil through a mud-water separation device to obtain mud water, and detects the density of the settled mud water through an intelligent density meter; the multi-parameter sensor unit (1) comprises a particle size analysis module (11), a moisture content detection module (12), a slag soil density detection module (13), a mud-water separation module (14) and a mud-water density detection module (15): The particle size analysis module (11) measures the particle size distribution of the slag soil particles in real time through a laser particle size analyzer, and obtains the particle size proportion information of the fine-grained soil in the slag soil; The moisture content detection module (12) detects the moisture content of the slag soil in real time through a moisture content sensor; The slag soil density detection module (13) detects and determines the density of the slag soil through a density sensor; The mud-water separation module (14) separates the slag soil through a mud-water separation device to obtain mud water; The mud-water density detection module (15) detects the density of the settled mud water through an intelligent density meter, and calculates the content of fine-grained soil in the mud water according to the corresponding relationship between the mud water density and the fine-grained soil content; The data acquisition and transmission unit (2) is used for collecting and processing the electrical signals or digital signals output by the multi-parameter sensor unit (1), and then transmitting the shield slag soil data to the data processing and analysis unit (3) in real time; The data processing and analysis unit (3) performs filtering and denoising processing on the shield slag soil data, introduces a correction algorithm based on the data difference law obtained through indoor test, corrects the fine-grained soil content data detected by the intelligent density meter, uses a multi-parameter fusion algorithm to comprehensively analyze a plurality of processed shield slag soil data, and calculates the content of fine-grained soil in the shield slag soil in combination with the weight of the slag soil and the coarse-grained grading parameters; the data processing and analysis unit (3) comprises a preliminary verification module (31), a data processing module (32), a data alignment module (33), a data correction module (34), a multi-parameter feature extraction module (35), a multi-parameter fusion algorithm operation module (36) and a data management module (37): The preliminary verification module (31) receives the data of the data acquisition and transmission unit (2) in real time through a preset communication interface protocol, and preliminarily verifies the integrity and format of the data during the receiving process; for the data with problems, a retransmission request is automatically sent to the data acquisition and transmission unit (2); The data processing module (32) uses a plurality of filtering algorithms to process a plurality of shield slag soil data; The data alignment module (33) uses time synchronization technology to calibrate and align the time stamps of a plurality of shield slag soil data of each sensor, so that all parameters remain consistent in the time dimension, forming a unified time series data set; The data correction module (34) corrects the fine-grained soil content data detected by the intelligent density meter by introducing a correction algorithm based on the data difference law obtained from indoor tests. The multi-parameter feature extraction module (35) extracts features from the preprocessed and corrected data, mines the feature information related to the fine-grained soil content of each parameter, calculates the mass of fine-grained soil in unit volume of slag based on the slag weight data, and calculates the proportion of coarse particles and fine-grained soil based on the coarse particle grading data, forming a multi-dimensional feature vector. The multi-parameter fusion algorithm operation module (36) uses a multi-parameter fusion algorithm to comprehensively analyze the extracted multi-dimensional feature vector and calculate the content of fine-grained soil in the shield slag. The data management module (37) uses a distributed database architecture to store the processed data, including raw sensor data, preprocessed data, correction results, fusion calculation results, and related feature parameters. It also regularly backs up data to local hard drives and cloud storage. The display unit (4) displays the calculation results of the fine-grained soil content of the slag, the real-time measurement data of each sensor, the weight of the slag, the coarse particle grading, and related construction parameters in real time. When the detected fine-grained soil content exceeds the pre-set reasonable range, the alarm immediately sends an audible and visual alarm signal.

2. The multi-parameter fusion-based real-time detection system for shield muck fine-grained soil content according to claim 1, characterized in that, In the data acquisition and transmission unit (2), the electrical or digital signals output by the multi-parameter sensor unit (1) are converted into a unified data format. High-precision A / D converters are used to digitize analog signals. The multi-sensor data is integrated through a data acquisition card, and the collected data is transmitted in real time to the data processing and analysis unit (3) using wireless communication technology.

3. The multi-parameter fusion-based real-time detection system for shield muck fine-grained soil content according to claim 2, characterized in that, In the data processing module (32), for continuous data output by the laser particle size analyzer, moisture content sensor, and density sensor, a median filter algorithm is used to remove transient pulse interference. For data output by the intelligent density meter, a Kalman filter algorithm is used to predict the expected value of the data by establishing a dynamic model and dynamically correcting the actual measurement value and the predicted value.

4. The multi-parameter fusion-based real-time detection system for shield muck fine-grained soil content according to claim 3, characterized in that, The data correction module (34) corrects the fine-grained soil content data detected by the intelligent density meter by introducing a correction algorithm based on the data difference law obtained from indoor tests, including the following steps: A1. During the indoor test phase, a large number of slag samples with different fine-grained soil content are collected, and the intelligent density meter detection method and the traditional laboratory sieving method are used for detection respectively, and a difference database of the detection results of the two methods is established; A2. In the correction algorithm, the real-time detection value of the intelligent density meter is taken as the input, the corresponding correction coefficient is calculated by querying the difference database and combining the linear regression model or BP neural network model, and the detection value is corrected; A3. Update the difference database regularly, include the latest field detection data and laboratory comparison data, and continuously optimize the parameters of the correction model.

5. The deduction method of the real-time detection system for the fine-grained soil content of shield muck based on multi-parameter fusion, which is applicable to the real-time detection system for the fine-grained soil content of shield muck based on multi-parameter fusion according to any one of claims 1-4, characterized in that, The following steps are included: Ⅰ. Establish a historical database: During the shield construction process, the system continuously collects and stores multi-parameter detection data, shield tunneling parameters, and geological condition information at different construction stages to establish a historical database; II. Data feature extraction: In-depth analysis of data in the historical database, extraction of characteristic parameters closely related to the change of fine-grained soil content, determination of which parameters have strong correlation with the fluctuation of fine-grained soil content under different stratum conditions through statistical analysis method; III. Data feature selection: Using data mining technology to mine the potential correlation features between parameters, using feature selection algorithm to select the most representative and most contributive feature parameter subset from the extracted features; IV. Construction of inference model: Based on the selected feature parameters, machine learning algorithm or time series analysis method is used to construct the fine-grained soil content inference model; V. Real-time inference and result output: In the process of shield construction, the current multi-parameter detection data, spoil weight, coarse-grained grading, shield tunneling parameters and geological condition information are input into the verified and optimized inference model, the model quickly predicts the change trend of shield spoil fine-grained soil content in the next 5-10 ring tunneling period according to the input data, and the inference result is output to the display unit (4) for display.

6. The deduction method of the multi-parameter fusion-based real-time detection system for shield muck fine-grained soil content according to claim 5, characterized in that, In step IV, the following steps are included: B1. Based on the selected feature parameters, machine learning algorithm or time series analysis method is used to construct the fine-grained soil content inference model; B2. In the process of building the model, part of the data in the historical database is used as the training set to optimize and adjust the parameters of the model; B3. The remaining part of the data in the historical database is used as the test set to verify the constructed inference model, the prediction performance of the model is evaluated by calculating the error index between the model prediction result and the actual fine-grained soil content data in the test set; B4. If the error of the model is large, further analyze the reason, take corresponding optimization measures, then train and verify the model again until the model reaches satisfactory prediction accuracy and stability.

Images