Multi-risk-source tunnel construction early warning method

Abstract

The invention discloses a multi-risk-source tunnel construction early warning method, relates to the field of tunnel construction early warning, and provides a multi-risk-source tunnel construction early warning method which comprises tunnel vault and horizontal convergence deformation prediction based on a time sequence model and a tunnel risk prediction model based on the time sequence model. According to the method, tunnel face images and construction site monitoring data are expanded into important data sources, a corresponding database is established, and a tunnel face joint extraction model, a tunnel deformation prediction model and a tunnel deformation risk assessment model are established by using algorithms such as image recognition, a neural network and machine learning. And a tunnel construction risk assessment method is formed. The invention aims to provide scientific basis and technical support for safety construction and risk management of tunnel engineering by using an informatization technology.

Description

Technical Field

[0001] The present invention relates to the field of tunnel construction early warning, and in particular to a multi-risk source tunnel construction early warning method. Background Art

[0002] In tunnel engineering, traditional risk assessment methods often begin with the consideration of risk factors, followed by the establishment of a risk assessment indicator system, the analysis and rating of risk consequences using qualitative or quantitative methods, and the final conclusion with the proposal of various risk control measures. However, establishing a risk assessment indicator system for highway tunnels involves numerous factors, many uncertainties, and some factors are difficult to quantify. Furthermore, most indicators are difficult to obtain, and their determination relies heavily on engineering experience. This creates a multitude of potential risk sources, making it difficult to effectively warn and prevent construction risks. Therefore, establishing an efficient and comprehensive risk identification and management mechanism is crucial for risk assessment during tunnel construction.

[0003] Because traditional risk assessment methods have significant limitations in tunnel construction and decision-making, the rapid development of computer technology has led to the gradual entry of artificial intelligence (AI) into the field of tunnel engineering. AI, encompassing various algorithms such as machine learning, neural networks, and image recognition, is a key tool for promoting the digitization, automation, informatization, and intelligentization of tunnel engineering.

[0004] However, most risk assessment models consider a limited number of indicators, most of which are static, and the acquisition of indicator data depends on the excavation of the tunnel face, which limits the applicability of risk assessment methods. Summary of the Invention

[0005] In response to the above-mentioned problems, the present invention aims to provide a multi-risk source tunnel construction early warning method, which uses geological survey information, tunnel face images and construction site monitoring information as data sources, introduces image recognition and machine learning methods, and introduces dynamic monitoring data into tunnel construction risk assessment to form a multi-risk source tunnel construction early warning method.

[0006] The main idea of ​​the technical solution adopted by the present invention is to combine artificial intelligence technology with traditional risk assessment methods, expand tunnel face images and construction site monitoring data into important data sources, establish a corresponding database, and use image recognition, neural networks, machine learning and other algorithms to build a tunnel face joint extraction model, a tunnel deformation prediction model and a tunnel deformation risk assessment model, and form a tunnel construction risk assessment method.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A multi-risk source tunnel construction early warning method, characterized by comprising:

[0009] Step 1: Establish a database for tunnel deformation prediction;

[0010] Step 2: Use the ZOA algorithm to optimize the XGBOOST regression model and construct a tunnel deformation prediction model;

[0011] Step 3: Input the data in the database established in step 1 into the tunnel deformation prediction model established in step 2 for training to obtain a trained tunnel deformation prediction model, and input the field monitoring data into the trained tunnel deformation prediction model for prediction to obtain a predicted value;

[0012] Step 4: Build a risk level database based on the predicted values ​​in step 3;

[0013] Step 5: Use the ZOA algorithm to optimize the XGBOOST classification model and construct a tunnel surrounding rock risk level prediction model;

[0014] Step 6: Input the data in the risk level database in step 4 into the tunnel surrounding rock risk level prediction model in step 5 to predict the risk level during the construction period.

[0015] Furthermore, the tunnel deformation prediction database in step 1 includes the following data:

[0016] Rock uniaxial compressive strength, rock integrity index, joint width, joint height, burial depth, monitoring time, vault settlement, and horizontal convergence.

[0017] Furthermore, the optimization process of step 2 is as follows:

[0018] The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST regression model to obtain a tunnel deformation prediction model.

[0019] Furthermore, the optimization process of step 5 is as follows:

[0020] The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST classification model to obtain a tunnel surrounding rock risk level prediction model.

[0021] The beneficial effects of the present invention are as follows: Compared with the existing technology, the present invention improves on the existing technology by combining artificial intelligence technology with traditional risk assessment methods. By expanding tunnel face images and construction site monitoring data into important data sources, a corresponding database is established. Using algorithms such as image recognition, neural networks, and machine learning, a tunnel face joint extraction model, a tunnel deformation prediction model, and a tunnel deformation risk assessment model are constructed, thereby forming a tunnel construction risk assessment method. This application aims to use information technology to provide a scientific basis and technical support for the safe construction and risk management of tunnel projects.

[0022] This application also optimizes the three parameters of the XGBOOST model, namely the maximum number of iterations, depth, and learning rate, to make the tunnel surrounding rock risk level prediction model more accurate and the prediction results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of the method for early warning of multi-risk source tunnel construction according to the present invention.

[0024] Figure 2 This is a database flow chart of the multi-risk source tunnel construction early warning method of the present invention.

[0025] Figure 3 This is the ZOA-XGBOOST algorithm process of the present invention.

[0026] Figure 4 This is a flow chart for constructing an indicator system for the multi-risk source tunnel construction early warning method of the present invention.

[0027] Figure 5 This is the training result of the tunnel deformation prediction model of the present invention.

[0028] Figure 6 This is the training result of the tunnel surrounding rock risk level prediction model of the present invention. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0030] See also Figures 1-6 This application discloses a multi-risk source tunnel construction early warning method, the method flow chart is as follows Figure 1 As shown in Figure 2, the early warning method for multi-risk source tunnel construction specifically includes:

[0031] Step 1: Establish a tunnel deformation prediction database;

[0032] Specifically, the process of establishing the database in step 1 is as follows Figure 2As shown in the figure, the first step is to establish a tunnel deformation prediction index system, which includes: rock uniaxial compressive strength, rock integrity index, joint width, joint height, burial depth, and monitoring time; the above indicators are output as field monitoring data, and a tunnel deformation prediction database is established based on arch settlement and horizontal convergence.

[0033] That is to say, the tunnel deformation prediction database in step 1 specifically includes: rock uniaxial compressive strength, rock integrity index, joint width, joint height, burial depth, monitoring time, vault settlement, and horizontal convergence.

[0034] Table 1 Tunnel deformation prediction database (part)

[0035]

[0036]

[0037] Step 2: Use the ZOA algorithm to optimize the XGBOOST regression model and construct a tunnel deformation prediction model;

[0038] Among them, the optimization process of step 2 is:

[0039] The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST regression model to obtain a tunnel deformation prediction model.

[0040] Specifically, the specific process of using the ZOA algorithm to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST model is as follows: Figure 3 As shown;

[0041] The ZOA algorithm is divided into two stages. ZOA algorithm initialization: x i,j =lb j +r×(ub j -lb j )

[0042] where x i,j For individuals, lb j is the lower bound of parameter optimization, ub j The upper bound of parameter optimization, r∈rand(0,1);

[0043] Phase 1: Foraging

[0044]

[0045] where I∈rand(1,2), is the new best position, PZ j The best position before.

[0046] Phase 2: Escape

[0047]

[0048] where I∈rand(1,2), is the new best position, PZ j The best position before.

[0049] Step 3: Input the data in the database established in step 1 into the tunnel deformation prediction model established in step 2 for training, and obtain the trained tunnel deformation prediction model. Input the field monitoring data into the trained tunnel deformation prediction model for prediction, and obtain the predicted value. The training results are as follows: Figure 5 As shown, Figure 5 The vertical axis is the measured value, and the horizontal axis is the output value of the model, that is, the predicted value. The closer the scatter points in the figure are to the red line, the better the fitting effect is. The blue line is the fitting line.

[0050] It should be noted that the predicted values ​​in step 3 need to be compared with the actual monitored values. The accuracy and feasibility of the tunnel deformation prediction regression model are verified by the error between the predicted and actual monitored values. If the error is small, it means that the tunnel deformation prediction regression model is accurate and feasible, and the predicted values ​​of the tunnel deformation prediction regression model are also accurate and feasible. The predicted values ​​can then be used as new data to expand the database in step 1. If the error is large, the predicted values ​​are directly discarded.

[0051] Step 4: Build a risk level database based on the predicted values ​​in step 3;

[0052] It should be noted that the risk level database is composed of a tunnel deformation risk assessment index system, which is composed of multiple indicators. The predicted value in step 3 is an indicator in the tunnel deformation risk assessment index system. In addition to the horizontal convergence and vault settlement indicators output in step 3, the risk level database also includes engineering geological factors and design and construction factors. Specifically, the specific contents of the tunnel deformation risk assessment index system are as follows: Figure 4 shown.

[0053] Table 2 lists some of the data in the tunnel construction deformation risk assessment database, along with the corresponding risk level for each data point. It should be noted that by inputting the database data into the tunnel surrounding rock risk level prediction model for training, the model weights are obtained. This allows predictions to be made simply by inputting the data. This means that the risk level can be automatically derived from the input data without the need for manual processing or the use of other tools.

[0054] Table 2 Tunnel construction deformation risk assessment database (partial)

[0055]

[0056] It should be noted that the risk level database in step 4 is mainly based on the horizontal and vault deformations, surrounding rock integrity coefficient, and tunnel design and construction factors predicted in step 3.

[0057] Step 5: Use the ZOA algorithm to optimize the XGBOOST classification model and construct a tunnel surrounding rock risk level prediction model;

[0058] Specifically, the optimization process of step 5 is:

[0059] The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST classification model to obtain a tunnel surrounding rock risk level prediction model.

[0060] It should be noted that the optimization process of step 5 is the same as that of step 2. Figure 3 The only difference is that the optimized parameters in step 2 are fed into the XGBOOST regression model for training, resulting in a ZOA-XGBOOST-based tunnel deformation prediction model; whereas in step 5, the optimized parameters are fed into the XGBOOST classification model for training, resulting in a ZOA-XGBOOST-based tunnel surrounding rock risk level prediction model.

[0061] Step 6: Input the data in the risk level database in step 4 into the tunnel surrounding rock risk level prediction model in step 5 to predict the risk level during the construction period. The training results of the tunnel surrounding rock risk level prediction model are as follows: Figure 6 shown. Figure 6 The horizontal axis is the number of samples, and the vertical axis is the risk level. If the positions of the solid circles and the hollow circles in the figure are consistent, it means that the model results are consistent with the actual results. If the positions are inconsistent, it means that the model results are misplaced.

[0062] Specifically, the tunnel construction deformation risk assessment index of the section to be predicted contained in the risk level database in step 4 is input into the tunnel surrounding rock risk level prediction model in step 5 to predict the risk level during the construction period.

[0063] Example

[0064] To verify the feasibility and effectiveness of the tunnel construction risk assessment method proposed in this study, a section on the left side of a tunnel currently under construction in the TJ-3 section of a highway project was selected for engineering application. The tunnel section data and construction design information obtained from geological survey data are shown in Table 3. The section was constructed using the circular excavation method with core soil reserved.

[0065] Table 3 Collected surrounding rock geological information

[0066]

[0067] The tunnel face images collected from the tunnel construction site at three sections, the K v They are 0.84 respectively.

[0068] The parameters required for tunnel deformation prediction, compiled from Tables 3 and 1, were input into the ZOA-XGBOOST model for tunnel deformation prediction. Tunnel deformation within 21 days was predicted based on the monitoring frequency. The prediction results of the ZOA-XGBOOST model are shown in Table 4.

[0069] Table 4 Tunnel deformation prediction model results

[0070]

[0071] Among them: 0~3d are the construction site monitoring values, 4~21d are the model prediction values

[0072] As shown in Table 4, the deformation rates of the arch settlement and horizontal convergence of the three sections on the 21st day were all less than 0.2 mm / d, which can be judged as basically stable and suitable for secondary lining construction.

[0073] The obtained K v The values ​​of the tunnel deformation risk, the crown settlement value, the horizontal convergence value, and the other parameters required for tunnel deformation risk assessment were input into the ZOA-XGBOOST model for tunnel deformation risk assessment to obtain the predicted risk levels of the three sections, as shown in Table 5.

[0074] Table 5 Results of tunnel deformation risk assessment model

[0075]

[0076] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-risk source tunnel construction early warning method, characterized in that: include: Step 1: Establish a database for tunnel deformation prediction; Step 2: Use the ZOA algorithm to optimize the XGBOOST regression model and construct a tunnel deformation prediction model; Step 3: Input the data in the database established in step 1 into the tunnel deformation prediction model established in step 2 for training to obtain a trained tunnel deformation prediction model, and input the field monitoring data into the trained tunnel deformation prediction model for prediction to obtain a predicted value; Step 4: Build a risk level database based on the predicted values ​​in step 3; Step 5: Use the ZOA algorithm to optimize the XGBOOST classification model and construct a tunnel surrounding rock risk level prediction model; Step 6: Input the data in the risk level database in step 4 into the tunnel surrounding rock risk level prediction model in step 5 to predict the risk level during the construction period.

2. A multi-risk source tunnel construction early warning method according to claim 1, characterized in that: The tunnel deformation prediction database in step 1 contains the following data: Rock uniaxial compressive strength, rock integrity index, joint width, joint height, burial depth, monitoring time, vault settlement, and horizontal convergence.

3. The multi-risk source tunnel construction early warning method according to claim 1 is characterized in that: The optimization process of step 2 is: The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST regression model to obtain a tunnel deformation prediction model.

4. The multi-risk source tunnel construction early warning method according to claim 1, characterized in that: The optimization process of step 5 is as follows: The ZOA algorithm is used to optimize the maximum number of iterations, depth, and learning rate parameters of the XGBOOST classification model to obtain a tunnel surrounding rock risk level prediction model.

Images