Intelligent surrounding rock grade forecasting method based on TSP-geological information fusion

Abstract

The invention relates to the field of tunnel advanced geological forecasting, and provides a surrounding rock level intelligent forecasting method based on TSP-geological information fusion, which comprises the following steps: firstly, extracting TSP trend characteristics, and carrying out spatial registration on the TSP trend characteristics and tunnel face mileage in geological information; secondly, an abnormal data screening rule is formulated according to TSP trend characteristics, and abnormal data is removed; on the basis, processed TSP trend characteristics and construction surface surrounding rock levels serve as model input, different construction surface categories are divided for modeling, TSP detection data distribution under different construction surface conditions is different, and a surrounding rock level forecasting model is established in a knowledge driving and data driving combined mode. And finally, the surrounding rock information in the design stage is added to carry out intelligent decision making with a TSP single forecasting result, so that surrounding rock level forecasting is realized. The method can effectively improve the prediction accuracy, and is of great significance to guarantee the life and property safety of tunnel construction.

Description

Technical Field

[0001] The present invention relates to the field of tunnel advanced geological prediction, and in particular to a surrounding rock level intelligent prediction method based on TSP-geological information fusion. Background Art

[0002] The prediction of surrounding rock level accompanies the whole process of tunnel excavation. Surrounding rock level is an extremely important comprehensive indicator in advanced geological prediction. Predicting surrounding rock level needs to consider many factors, including the use of geological survey and design data, the interpretation of various geophysical data, and the analysis of actual groundwater conditions. Since it involves various multi-source heterogeneous detection data, it brings great challenges to the intelligent prediction of surrounding rock level. The traditional method of predicting surrounding rock level is to make judgments by professionals based on specific detection data combined with experience and professional knowledge. This method is highly subjective and requires high costs and low efficiency. With the advancement of technology, more and more researchers have established intelligent prediction models through neural network technology to predict surrounding rock level. However, the current prediction model still has many defects. 1. Due to the multi-source heterogeneous characteristics of detection data, it is difficult to extract effective features and the data quality is general; 2. The single prediction method based on TSP has multiple solutions, resulting in insufficient prediction accuracy. Therefore, in order to ensure the safety and efficiency of tunnel construction, it is urgent to study a surrounding rock level intelligent prediction method based on TSP-geological information fusion. Summary of the invention

[0003] In order to solve the problems of coexistence of multi-source heterogeneous data in the existing tunnel advanced geological prediction detection data, unclear differences in detection data characteristics, resulting in difficulty in effectively predicting the surrounding rock grade, and multiple solutions of a single prediction method, the present invention provides a surrounding rock grade intelligent prediction method based on TSP-geological information fusion, which improves the quality of TSP detection data and the accuracy of surrounding rock grade prediction by extracting features from TSP detection data and screening data according to expert knowledge, considering using TSP detection data to establish different sub-models according to the layered method of construction surface and incorporating surrounding rock information in the design stage for intelligent decision-making, which is of great significance to ensuring the safety and efficiency of tunnel construction.

[0004] The technical solution of the present invention is as follows:

[0005] An intelligent prediction method for surrounding rock level based on TSP-geological information fusion, comprising the following steps:

[0006] S1: The TSP trend characteristics are obtained by using the feature extraction method of the applied surface difference and then normalized;

[0007] S2: Obtain different mileage points of the tunnel face in the geological information, and perform spatial registration with the TSP trend features obtained in S1 to form a data set of TSP trend features and surrounding rock levels. For the obtained data set, formulate TSP trend feature abnormal data screening rules to eliminate abnormal data that do not meet the requirements;

[0008] S3: According to the feature extraction and abnormal data screening method, the effective feature data corresponding to each tunnel face is obtained; the surrounding rock level of the construction face is stratified, and a TSP surrounding rock level prediction sub-model is established based on the trend characteristics of the construction face surrounding rock level and the construction face surrounding rock level information according to the stratified levels by combining knowledge-driven and data-driven methods. Finally, the surrounding rock information in the design stage is incorporated into the TSP surrounding rock level prediction sub-model to form a TSP-geology fusion model for intelligent decision-making;

[0009] S4: Use the trained TSP surrounding rock level prediction sub-model and TSP-geology fusion model to make advanced geological predictions for actual tunnel projects and verify the models.

[0010] Furthermore, step S1 specifically includes the following process:

[0011] S1.1: Extract the trend characteristics of TSP detection data as follows:

[0012] V_trend=V-V0 (1.1)

[0013] Among them, V_trend is the trend feature of TSP data, V is the original data of TSP detection for each point, and V0 is the TSP detection data of the working surface of this TSP detection;

[0014] S1.2: Then the extracted trend features are standardized to obtain the final TSP trend feature library. The formula for the standardized processing is as follows:

[0015]

[0016] Where Z is the standardized processing result of TSP trend characteristics, V_trend is the trend characteristics extracted from TSP detection data, υ is the mean of TSP trend characteristics, and σ is the standard deviation of TSP trend characteristics.

[0017] Furthermore, step S2 specifically includes the following process:

[0018] S2.1: Spatially align the trend characteristics of the TSP detection data with the tunnel face mileage as follows:

[0019]

[0020] Where S′ is the tunnel mileage aligned with the actual face mileage, SZZM is the actual tunnel face mileage, S i is the tunnel mileage corresponding to the ith TSP trend characteristic data, L is the tunnel excavation direction, 0 is a small mileage, and 1 is a large mileage;

[0021] S2.2: For TSP trend characteristic data, when using it for advanced geological forecasting, if the surrounding rock level of the construction surface has been determined, when it rises relative to the construction surface, but the actual surrounding rock condition is worse, when it falls relative to the construction surface, but the actual surrounding rock condition is better, such trend characteristic data that do not conform to the mechanism are eliminated. The specific mechanism meets the following conditions:

[0022] △V p =V p -V p0 (1.8)

[0023]

[0024] Among them, V p is the longitudinal wave velocity at a certain mileage point, V p0 is the longitudinal wave velocity of the construction surface, ΔV p is the change in longitudinal wave velocity at a certain mileage point relative to the construction surface, F TSP_data Represents the TSP abnormal trend characteristic data that has been eliminated, B is the actual surrounding rock grade, and A is the surrounding rock grade of the construction surface.

[0025] Furthermore, step S3 specifically includes the following process:

[0026] S3.1: According to the feature matrix extraction and abnormal data screening method, the TSP trend characteristics of each tunnel face are obtained, and a data set is formed and divided into a training set, a validation set, and a test set;

[0027] S3.2: The surrounding rock grades of the construction surface are divided into Grade I, Grade II, Grade III, Grade IV, Grade V and Grade VI. The TSP trend characteristic matrix and surrounding rock grades are used to build models. For construction surfaces that can be directly divided, expert experience is used for judgment. For those that cannot be directly divided, data-driven modeling is used to build models, and the TSP surrounding rock grade prediction sub-model is obtained.

[0028] S3.3: According to the TSP surrounding rock level prediction sub-model established on different construction surfaces, a surrounding rock result is obtained by prediction. Then, the surrounding rock information in the design stage is incorporated into the prediction result of the TSP surrounding rock level model to make an intelligent decision, and a surrounding rock level prediction model based on TSP-geology fusion is obtained. The specific decision-making mechanism is as follows:

[0029] If the TSP forecast result is too different from the surrounding rock grade in the design stage, the TSP forecast result shall prevail; if the TSP forecast result is not much different from the surrounding rock grade in the design stage, different weights shall be considered for comprehensive decision-making. The weights of the surrounding rock grade in the design stage and the TSP forecast model are 0.4 and 0.6, respectively.

[0030] Furthermore, step S4 specifically includes the following process:

[0031] S4.1: Use the TSP surrounding rock level prediction sub-model and TSP-geology fusion model to predict the test set data in S5 and verify the effect of the model;

[0032] S4.2: Use the test set for verification. The verification indicators are as follows:

[0033]

[0034] Where Acc is the test set accuracy, NUM True is the number of correctly predicted samples, NUM Total is the total number of samples in the test set.

[0035] The beneficial effects brought by the technical solution provided by the present invention are as follows:

[0036] The present invention proposes a method for extracting trend features from TSP detection data, and in order to improve data quality, an outlier screening method for TSP trend features is proposed based on expert experience; considering that a single prediction method has multiple solutions and it is difficult to accurately predict the surrounding rock level in front of the face, an intelligent prediction method for surrounding rock level based on TSP-geological information fusion is proposed. First, the surrounding rock level is stratified according to different working faces, and a surrounding rock level prediction model of TSP detection data and different working face geological conditions is established by combining knowledge-driven and data-driven methods. Then, the surrounding rock information in the design stage is considered to be integrated for intelligent decision-making to achieve accurate prediction of the surrounding rock level, and the prediction effect of the model is simulated and verified using test set data, which is conducive to the application of the present invention in actual production. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is an architecture diagram of a surrounding rock level intelligent prediction method based on TSP-geological information fusion in an embodiment of the present invention.

[0038] Figure 2 It is the test result of the TSP sub-model and the TSP-geology fusion prediction model in the embodiment of the present invention. DETAILED DESCRIPTION

[0039] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, the technical solutions of the present invention are now described in detail with reference to the accompanying drawings and embodiments.

[0040] Figure 1 This is the architecture diagram of the surrounding rock level intelligent prediction method based on TSP-geological information fusion provided by the present invention. The surrounding rock level intelligent prediction method is divided into three parts. The first part performs feature extraction, outlier screening, spatial registration and other operations on the acquired TSP detection data to form a data set of TSP trend feature data and surrounding rock level; the second part first stratifies according to the different surrounding rock levels of the construction surface, and uses expert experience and convolutional neural network to establish a model for TSP detection data to predict surrounding rock level; the third part incorporates the surrounding rock information in the design stage and the surrounding rock level predicted by the TSP model to assist in decision-making and realize the fusion prediction of surrounding rock level. Finally, the TSP sub-model and the TSP-geological fusion prediction model are simulated and verified using the test set data.

[0041] Example:

[0042] An intelligent prediction method for surrounding rock level based on TSP-geological information fusion, including:

[0043] S1: For TSP detection data, the interpretation of TSP detection data is to make a forecast through the rising or falling trend of TSP data of the forecast point and the construction surface. In this embodiment, the feature extraction method of the construction surface difference is used to obtain the TSP trend feature and perform normalization processing;

[0044] S1.1: For structured TSP detection data, in actual engineering sites, geophysical prospectors use the degree of rise or fall of TSP detection data relative to the working face (the tunnel face where geophysical exploration is conducted) to make advance geological predictions. This paper considers extracting trend characteristics of TSP detection data, as shown below:

[0045] V_trend=V-V0 (1.1)

[0046] Among them, V_trend is the trend feature of TSP data, V is the original data of TSP detection for each point, and V0 is the TSP detection data of the working surface of this TSP detection;

[0047] S1.2: Then the extracted trend features are standardized to obtain the final TSP trend feature library. The formula for the standardization is as shown in 1-2.

[0048]

[0049] Where Z is the standardized processing result of TSP trend characteristics, V_trend is the trend characteristics extracted from TSP detection data, υ is the mean of TSP trend characteristics, and σ is the standard deviation of TSP trend characteristics.

[0050] S2: For different tunnel face mileage points obtained from geological information, spatial registration is performed with the TSP trend features obtained in S1. Since the surrounding rock grade information of each tunnel face mileage point is known, the registration of the surrounding rock grade and the TSP trend features can be completed to form a data set of TSP trend features and surrounding rock grade; then, for the obtained data set, in order to improve the data quality, a TSP trend feature abnormal data screening rule is formulated to eliminate abnormal data that does not meet the requirements;

[0051] S2.1: For the trend feature of TSP detection data, the corresponding tunnel mileage is the detection mileage, so it is necessary to first align it with the face mileage in space. This embodiment uses the nearest neighbor interpolation method to achieve this. According to actual tunnel engineering experience, this method is related to the tunnel excavation direction. For example, for a tunnel excavated from a small mileage to a large mileage, the face mileage should correspond to the mileage that is smaller than it and closest to it. Based on this, the mileage alignment can be completed, that is, the registration of the TSP trend feature and the surrounding rock level is completed, as shown below:

[0052]

[0053] Where S′ is the tunnel mileage aligned with the actual face mileage, S ZZM is the actual tunnel face mileage, S i is the tunnel mileage corresponding to the ith TSP trend characteristic data, and L is the tunnel excavation direction (0 is a small mileage and 1 is a large mileage).

[0054] S2.2: For TSP trend feature data, when using it for advanced geological forecasting, if the surrounding rock level of the construction surface has been determined, when it rises relative to the construction surface, but the actual surrounding rock condition is worse, when it falls relative to the construction surface, but the actual surrounding rock condition is better, this type of trend feature data does not conform to the mechanism, so this embodiment considers adding a mechanism to eliminate this type of data, so as to improve data quality. The specific mechanism meets the following conditions:

[0055] ΔV p =V p -V p0 (1.8)

[0056]

[0057] Among them, V p is the longitudinal wave velocity at a certain mileage point, V p0 is the longitudinal wave velocity of the construction surface, ΔV p is the change in longitudinal wave velocity at a certain mileage point relative to the construction surface, F TSP_data Represents the TSP abnormal trend characteristic data that has been eliminated, B is the actual surrounding rock grade, and A is the surrounding rock grade of the construction surface.

[0058] S3: According to the feature extraction and abnormal data screening method, the effective feature data corresponding to each tunnel face is obtained; considering the TSP trend characteristics and the surrounding rock grade information of the construction face, a surrounding rock grade prediction model is established, and finally the surrounding rock information of the design stage is incorporated and intelligent decision-making is performed with the prediction results of the TSP sub-model;

[0059] S3.1: According to the feature matrix extraction and abnormal data screening method, the TSP trend characteristics of each tunnel face are obtained, and a data set is formed and divided into a training set, a validation set, and a test set;

[0060] S3.2: The surrounding rock grades of the construction surface are stratified into Grade I, Grade II, Grade III, Grade IV, Grade V and Grade VI. The TSP trend characteristic matrix and surrounding rock grade are used to build a model. For the construction surfaces that can be directly divided, expert experience is used for judgment. For those that cannot be directly divided, a data-driven approach is used for modeling to obtain the final surrounding rock grade prediction model.

[0061] S3.3: According to the TSP surrounding rock grade prediction model established on different construction surfaces, a surrounding rock result is obtained; then the surrounding rock information in the design stage is incorporated into the prediction result of the TSP surrounding rock grade model for intelligent decision-making, and a surrounding rock grade prediction model based on TSP-geology fusion is obtained. The specific decision-making mechanism is as follows: if the TSP prediction result is too different from the surrounding rock grade in the design stage, the TSP prediction result shall prevail; if the TSP prediction result is not much different from the surrounding rock grade in the design stage, different weights are considered for comprehensive decision-making. The weights of the surrounding rock grade in the design stage and the TSP prediction model are 0.4 and 0.6, respectively.

[0062] S4: Use the trained tunnel surrounding rock level intelligent prediction model to make advanced geological predictions for actual tunnel projects and verify the model.

[0063] S4.1: Use the TSP sub-model and TSP-geological fusion model to predict the test set data in S5 and verify the effect of the model;

[0064] S4.2: Use the test set for verification. The verification indicators are as follows:

[0065]

[0066] Where Acc is the test set accuracy, NUM True is the number of correctly predicted samples, NUM Total is the total number of samples in the test set.

[0067] The TSP surrounding rock level sub-model and TSP-geology fusion prediction model were verified using the test set data. Figure 2Shown are the accuracy results of the surrounding rock level prediction model. The results prove that the use of the surrounding rock level intelligent prediction method based on TSP-geological information fusion proposed in the present invention can improve the accuracy of single TSP prediction and reduce the multi-solution of single TSP prediction, which is conducive to the application of the present invention in actual production.

[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent prediction method for surrounding rock level based on TSP-geological information fusion, characterized in that: The steps include: S1: The TSP trend characteristics are obtained by using the feature extraction method of the applied surface difference and then normalized; S2: Obtain different mileage points of the tunnel face in the geological information, and perform spatial registration with the TSP trend features obtained in S1 to form a data set of TSP trend features and surrounding rock levels. For the obtained data set, formulate TSP trend feature abnormal data screening rules to eliminate abnormal data that do not meet the requirements; S3: According to the feature extraction and abnormal data screening method, the effective feature data corresponding to each tunnel face is obtained; the surrounding rock grade of the construction face is stratified, and a TSP surrounding rock grade prediction sub-model is established based on the trend characteristics of the construction face surrounding rock grade and the construction face surrounding rock grade information according to the stratified levels by combining knowledge-driven and data-driven methods. Finally, the surrounding rock information in the design stage is incorporated into the TSP surrounding rock grade prediction sub-model to form a TSP-geology fusion model for intelligent decision-making; S4: Use the trained TSP surrounding rock level prediction sub-model and TSP-geology fusion model to make advanced geological predictions for actual tunnel projects and verify the models.

2. The method for intelligent prediction of surrounding rock level based on TSP-geological information fusion according to claim 1, characterized in that: Step S1 specifically includes the following process: S1.1: Extract the trend characteristics of TSP detection data as follows: V_trend=V-V0 (1.1) Among them, V_trend is the trend feature of TSP data, V is the original data of TSP detection for each point, and v0 is the TSP detection data of the working surface of this TSP detection; S1.2: Then the extracted trend features are standardized to obtain the final TSP trend feature library. The formula for the standardized processing is as follows: Where Z is the standardized processing result of TSP trend characteristics, V_trend is the trend characteristics extracted from TSP detection data, v is the mean of TSP trend characteristics, and σ is the standard deviation of TSP trend characteristics.

3. The method for intelligent prediction of surrounding rock level based on TSP-geological information fusion according to claim 2, characterized in that: Step S2 specifically includes the following process: S2.1: Spatially align the trend characteristics of the TSP detection data with the tunnel face mileage as follows: Where S′ is the tunnel mileage aligned with the actual face mileage, S ZZM is the actual tunnel face mileage, S i is the tunnel mileage corresponding to the ith TSP trend characteristic data, L is the tunnel excavation direction, 0 is a small mileage, and 1 is a large mileage; S2.2: For TSP trend characteristic data, when using it for advanced geological forecasting, if the surrounding rock level of the construction surface has been determined, when it rises relative to the construction surface, but the actual surrounding rock condition is worse, when it falls relative to the construction surface, but the actual surrounding rock condition is better, such trend characteristic data that do not conform to the mechanism are eliminated. The specific mechanism meets the following conditions: ΔV p =V p -V p0 (1.8) Among them, V p is the longitudinal wave velocity at a certain mileage point, V p0 is the longitudinal wave velocity of the construction surface, ΔV p is the change in longitudinal wave velocity at a certain mileage point relative to the construction surface, F TSP_data Represents the TSP abnormal trend characteristic data that has been eliminated, B is the actual surrounding rock grade, and A is the surrounding rock grade of the construction surface.

4. The method for intelligent prediction of surrounding rock level based on TSP-geological information fusion according to claim 3, characterized in that: Step S3 specifically The process includes: S3.1: According to the feature matrix extraction and abnormal data screening method, the TSP trend characteristics of each tunnel face are obtained, and a data set is formed and divided into a training set, a validation set, and a test set; S3.2: The surrounding rock grades of the construction surface are divided into Grade I, Grade II, Grade III, Grade IV, Grade V and Grade VI. The TSP trend characteristic matrix and surrounding rock grades are used to build models. For construction surfaces that can be directly divided, expert experience is used for judgment. For those that cannot be directly divided, data-driven modeling is used to build models, and the TSP surrounding rock grade prediction sub-model is obtained. S3.3: According to the TSP surrounding rock level prediction sub-model established on different construction surfaces, a surrounding rock result is obtained by prediction. Then, the surrounding rock information in the design stage is incorporated into the prediction result of the TSP surrounding rock level model to make an intelligent decision, and a surrounding rock level prediction model based on TSP-geology fusion is obtained. The specific decision-making mechanism is as follows: If the TSP forecast result is too different from the surrounding rock grade in the design stage, the TSP forecast result shall prevail; if the TSP forecast result is not much different from the surrounding rock grade in the design stage, different weights shall be considered for comprehensive decision-making. The weights of the surrounding rock grade in the design stage and the TSP forecast model are 0.4 and 0.6, respectively.

5. The method for intelligent prediction of surrounding rock level based on TSP-geological information fusion according to claim 4, characterized in that: Step S4 specifically includes the following process: S4.1: Use the TSP surrounding rock level prediction sub-model and TSP-geology fusion model to predict the test set data in S5 and verify the effect of the model; S4.2: Use the test set for verification. The verification indicators are as follows: Where Acc is the test set accuracy, NUM True is the number of correctly predicted samples, NUM Total is the total number of samples in the test set.

Images