Nothing Special   »   [go: up one dir, main page]

CN112287595B - Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning - Google Patents

Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning Download PDF

Info

Publication number
CN112287595B
CN112287595B CN202010975476.2A CN202010975476A CN112287595B CN 112287595 B CN112287595 B CN 112287595B CN 202010975476 A CN202010975476 A CN 202010975476A CN 112287595 B CN112287595 B CN 112287595B
Authority
CN
China
Prior art keywords
grouting
ground penetrating
penetrating radar
shield tunnel
steps
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010975476.2A
Other languages
Chinese (zh)
Other versions
CN112287595A (en
Inventor
谢雄耀
刘凤洲
曾里
周彪
石锦江
刘浩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tongji University
Jinan Rail Transit Group Co Ltd
Original Assignee
Tongji University
Jinan Rail Transit Group Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tongji University, Jinan Rail Transit Group Co Ltd filed Critical Tongji University
Priority to CN202010975476.2A priority Critical patent/CN112287595B/en
Publication of CN112287595A publication Critical patent/CN112287595A/en
Application granted granted Critical
Publication of CN112287595B publication Critical patent/CN112287595B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH OR ROCK DRILLING; MINING
    • E21D—SHAFTS; TUNNELS; GALLERIES; LARGE UNDERGROUND CHAMBERS
    • E21D11/00—Lining tunnels, galleries or other underground cavities, e.g. large underground chambers; Linings therefor; Making such linings in situ, e.g. by assembling
    • E21D11/04—Lining with building materials
    • E21D11/10—Lining with building materials with concrete cast in situ; Shuttering also lost shutterings, e.g. made of blocks, of metal plates or other equipment adapted therefor
    • E21D11/105—Transport or application of concrete specially adapted for the lining of tunnels or galleries ; Backfilling the space between main building element and the surrounding rock, e.g. with concrete
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
    • G01V3/12—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with electromagnetic waves
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Geometry (AREA)
  • Structural Engineering (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Architecture (AREA)
  • Electromagnetism (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Civil Engineering (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Geology (AREA)
  • Data Mining & Analysis (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Mining & Mineral Resources (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Geophysics (AREA)
  • Medical Informatics (AREA)
  • Development Economics (AREA)
  • Software Systems (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Environmental & Geological Engineering (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)

Abstract

The invention discloses a prediction method of post grouting thickness of a shield tunnel wall based on ground penetrating radar detection and machine learning, which utilizes an XGboost principle to simulate post grouting of the shield tunnel wall in a model test and collect data to construct a data set, constructs an XGBoost model after data preprocessing, and predicts and identifies a ground penetrating radar image acquired by real-time detection of the post grouting ground penetrating radar of the shield tunnel wall by writing a prediction module. The method has the advantages that the method is more attached to the actual condition of the post grouting of the shield tunnel wall, and the ground penetrating radar image of the post grouting of the shield tunnel wall can be effectively predicted, so that the possible dangerous situation can be predicted in advance.

Description

Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning
Technical Field
The invention relates to the field of shield tunnel wall post-grouting detection, in particular to a prediction method of shield tunnel wall post-grouting thickness based on ground penetrating radar detection and machine learning.
Background
The existing detection of grouting behind the shield tunnel wall mainly takes the detection of a ground penetrating radar as a main part, and the thickness and uniformity of grouting are detected through the ground penetrating radar device. However, the image of the ground penetrating radar is not a direct image of the underground structure, and the return signal is an image formed by electromagnetic wave signals, and analysis and explanation are required. And (5) checking the tunnel site by a detector, and recording and identifying the ground penetrating radar image of grouting behind the shield tunnel wall. The detection method is low in detection speed, extremely strong in experience dependence on detection personnel and high in subjectivity. Therefore, a rapid, effective and accurate detection method for radar images based on machine learning is urgently needed. This is where the present application requires significant improvement.
Disclosure of Invention
The invention aims to provide a prediction method of post grouting thickness of a shield tunnel wall based on ground penetrating radar detection and machine learning.
In order to solve the technical problems, the invention provides a prediction method of the post grouting thickness of a shield tunnel wall based on ground penetrating radar detection and machine learning, which comprises the following steps:
s1, acquiring signals and images in a shield tunnel by using a ground penetrating radar, and corresponding the ground penetrating radar signals to scanning positions to form a data set;
s2, simulating a model test of grouting behind a shield tunnel wall of the ground penetrating radar, acquiring a ground penetrating radar image of the shield tunnel with known grouting thickness, and constructing a data set;
s3, after preprocessing the data, establishing a prediction model of XGboost grouting thickness based on the ground penetrating radar image according to the characteristics of the data set;
s4, optimizing parameters of the prediction model of the XGboost grouting thickness;
s5, predicting a ground penetrating radar image acquired by the grouting ground penetrating radar in real time after the shield tunnel wall by using a prediction model of the XGboost grouting thickness after parameter optimization.
The step S1 includes the steps of:
s11, using a ground penetrating radar to scan and detect lining of the tail of a shield tunnel and grouting behind the wall of the shield tunnel under construction, wherein the ground penetrating radar is positioned close to the duct piece;
s12, recording the scanning position and the scanning path of the ground penetrating radar, collecting signals of the ground penetrating radar, and marking the corresponding position of each signal on the scanning path;
s13, numbering each acquired radar signal from 0;
and S14, corresponding the ground penetrating radar signal to the scanning position to form a data set.
The step S2 includes the steps of:
s21, selecting shield tunnel segments and grouting slurry for model test, and according to the distribution 1 of the segments and grouting: 1, making a model, placing a duct piece in the same soil, injecting the same slurry, and placing a wave-absorbing material under boundary condition treatment;
s22, the thickness of the grouting layer in the model is divided into three types of undergrouting, normal grouting and over grouting according to engineering ground conditions and engineering experience;
s23, performing a model test by using the ground penetrating radar, and scanning grouting layers with known thickness classifications;
s24, recording the scanning position and scanning path of the radar, collecting signals of the ground penetrating radar, marking the corresponding position of each signal on the scanning path, and dividing the signals into three types of undergrouting, normal grouting and over grouting.
The step S3 includes the steps of:
s31, carrying out data preprocessing on radar signals acquired in a model test and a shield tunnel, wherein the data preprocessing comprises the steps of drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization in sequence, and the obtained radar signals are data between [ -1,1] and form a sample set;
s32, dividing a sample set into a training sample set I, a verification sample set and a test sample set according to a random mode for radar data collected by a preprocessed model test;
s33, calculating a loss function of the training sample set I in the current round according to the training sample set I:
Figure 831636DEST_PATH_IMAGE001
；
in the method, in the process of the invention,
Figure 387382DEST_PATH_IMAGE002
for loss function->
Figure 516881DEST_PATH_IMAGE003
For training sample number, ++>
Figure 892499DEST_PATH_IMAGE004
For->
Figure 844274DEST_PATH_IMAGE005
Training samples
Figure 739680DEST_PATH_IMAGE006
Loss of->
Figure 56392DEST_PATH_IMAGE007
For training the true tag value of the sample, +.>
Figure 168573DEST_PATH_IMAGE008
Is the firstiThe first sample is att-1Strong learner prediction value at multiple iterations,/->
Figure 533827DEST_PATH_IMAGE009
For the ith training sample->
Figure 558108DEST_PATH_IMAGE010
In the first placetA weak learner function during the training of the second iteration,/->
Figure 780141DEST_PATH_IMAGE011
And->
Figure 458247DEST_PATH_IMAGE012
For manually set coefficients +.>
Figure 673197DEST_PATH_IMAGE013
For the number of leaf nodes, +.>
Figure 58042DEST_PATH_IMAGE014
Is the first leaf node value;
s34, calculating an experience loss function of the current sample according to the training sample set I
Figure 201710DEST_PATH_IMAGE015
First and second partial derivatives based on a previous machine learning, and summing the first and second partial derivatives:
Figure 976899DEST_PATH_IMAGE016
；
in the method, in the process of the invention,
Figure 729960DEST_PATH_IMAGE017
for loss function->
Figure 234891DEST_PATH_IMAGE018
To fall into->
Figure 546530DEST_PATH_IMAGE019
Training sample set of individual leaf nodes, +.>
Figure 605753DEST_PATH_IMAGE020
Is->
Figure 162505DEST_PATH_IMAGE021
Loss function pair of individual samples +.>
Figure 256363DEST_PATH_IMAGE021
First partial derivative of the predicted value of the individual samples, ->
Figure 318997DEST_PATH_IMAGE022
Is->
Figure 616248DEST_PATH_IMAGE021
Loss function pair of individual samples +.>
Figure 727424DEST_PATH_IMAGE021
Second partial derivative of the predicted value of the individual samples,>
Figure 925056DEST_PATH_IMAGE023
and->
Figure 299536DEST_PATH_IMAGE024
For manually set coefficients, empirical loss functionsLIs part of the loss function, i.e., the remainder of the regularization is removed;
s35, the number of passing wheels istThe ratio of =1, 2, the iteration of T, and establishing an XGboost prediction model.
The step S35 includes the steps of:
s351, calculate the first
Figure 661248DEST_PATH_IMAGE025
The individual samples are in the empirical loss function->
Figure 323917DEST_PATH_IMAGE026
Based on->
Figure 330050DEST_PATH_IMAGE027
Is>
Figure 921437DEST_PATH_IMAGE028
And second derivative->
Figure 770444DEST_PATH_IMAGE029
Calculate the sum of the first derivatives of all samples +.>
Figure 489002DEST_PATH_IMAGE030
Sum of second derivatives of all samples +.>
Figure 365953DEST_PATH_IMAGE031
，
Figure 206870DEST_PATH_IMAGE032
=1，2，…，
Figure 418540DEST_PATH_IMAGE033
；
S352, based on the current node attempting to split the decision tree, for all featureskCalculating a maximum score;
s353 based on maximum
Figure 3105DEST_PATH_IMAGE034
Splitting subtrees by corresponding dividing features and feature values;
s354, if maximum
Figure 295415DEST_PATH_IMAGE034
If the value is 0, the current decision tree is built, and the +.>
Figure 182599DEST_PATH_IMAGE035
，
Figure 271778DEST_PATH_IMAGE035
Is the firsttWheel training firstjThe values of the individual leaf nodes, obtaining weak learner->
Figure 752045DEST_PATH_IMAGE036
Updating strong learner->
Figure 977490DEST_PATH_IMAGE037
Enter the next round of weak learner iteration if max +>
Figure 97892DEST_PATH_IMAGE038
Not 0, go to step S352 to continueAn attempt is made to split the decision tree.
The step S352 includes the steps of:
s3521 making the initial maximum score
Figure 533422DEST_PATH_IMAGE039
，
Figure 459790DEST_PATH_IMAGE040
、
Figure 415107DEST_PATH_IMAGE041
；
S3522 sample is characterized by
Figure 768728DEST_PATH_IMAGE042
Sequentially taking out the +.>
Figure 989756DEST_PATH_IMAGE043
And sequentially calculating the sum of the first derivatives and the second derivatives of the left subtree and the right subtree after the current sample is placed in the left subtree:
Figure 595181DEST_PATH_IMAGE044
，
Figure 795218DEST_PATH_IMAGE045
；
Figure 178795DEST_PATH_IMAGE046
，
Figure 136387DEST_PATH_IMAGE047
；
in the method, in the process of the invention,
Figure 404557DEST_PATH_IMAGE048
representing a numerical update->
Figure 640909DEST_PATH_IMAGE049
Is left toSum of all sample first derivatives of subtrees, < +.>
Figure 601912DEST_PATH_IMAGE050
Sum of all samples first derivatives of right subtree, +.>
Figure 46800DEST_PATH_IMAGE051
For the sum of the first derivatives of all samples, +.>
Figure 977715DEST_PATH_IMAGE052
Sum of all sample first derivatives of left subtree,/->
Figure 152345DEST_PATH_IMAGE053
Is the sum of the second derivatives of all samples of the right subtree, < ->
Figure 628457DEST_PATH_IMAGE054
Sum of second derivatives for all samples;
Figure 888537DEST_PATH_IMAGE055
And->
Figure 921346DEST_PATH_IMAGE056
The first derivative and the second derivative of the ith sample of the t-th round entering the left subtree are respectively;
s3523, updating the maximum score:
Figure 825848DEST_PATH_IMAGE057
。
the step S4 includes the steps of:
s41, introducing K-fold cross verification when the XGBoost algorithm is trained, estimating a generalization error of a classifier, and evaluating the classification accuracy of the XGBoost algorithm under different parameters when only a training sample is possessed;
s42, searching the multidimensional array in parallel from different growth directions by adopting a grid search algorithm, and determining parameters of grid search.
The step S42 includes the steps of:
s421, determining a search range of parameters to be searched according to experience;
s422, setting a search step length of the searched parameter;
s423, calculating the classification accuracy of XGBoost according to a cross verification method for each group of values on the grid;
and S424, giving the classification accuracy of each group, and determining the optimal parameter value according to the classification accuracy.
The step S5 includes the steps of:
s51, preprocessing the ground penetrating radar data acquired in the S1, wherein the steps of data preprocessing sequentially comprise drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization, and the obtained radar signals are data between [ -1,1 ];
s52, inputting the data into an established XGboost prediction model to obtain a prediction value of the grouting layer thickness classification.
The invention has the advantages that: the machine learning prediction method applied to the field of grouting detection behind the shield tunnel wall utilizes the XGboost principle to quickly, effectively and accurately identify the ground penetrating radar image; compared with the prior art, the prediction method is more attached to the actual condition of grouting behind the shield tunnel wall, and can effectively predict the ground penetrating radar image of grouting behind the shield tunnel wall, so that dangerous situations possibly occurring are predicted in advance, precious time is provided for timely supplementing grouting, the problem that grouting data behind the shield tunnel wall are difficult to interpret is effectively solved, and real-time feedback and guidance are provided for grouting construction.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention. In the drawings:
FIG. 1 is a flow chart of the invention S1-S5;
FIG. 2 is a flowchart illustrating the XGBoost model of S3;
FIG. 3 is a diagram illustrating the grid parameter search in S4 of the present invention;
fig. 4 is a flowchart illustrating the operation of the grid parameter search in S4 of the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Fig. 1 shows a flow chart of an embodiment of the invention. As shown in fig. 1-4, the invention provides a prediction method of post grouting thickness of a shield tunnel wall based on ground penetrating radar detection and machine learning, which comprises the following steps:
s1, acquiring signals and images in a shield tunnel by using a ground penetrating radar, and corresponding the ground penetrating radar signals to scanning positions to form a data set, wherein the method specifically comprises the following steps of:
s11, using a ground penetrating radar to scan and detect lining of the tail of a shield tunnel and grouting behind the wall of the shield tunnel under construction, wherein the ground penetrating radar is positioned close to the duct piece;
s12, recording the scanning position and the scanning path of the ground penetrating radar, collecting signals of the ground penetrating radar, and marking the corresponding position of each signal on the scanning path;
s13, numbering each acquired radar signal from 0;
and S14, corresponding the ground penetrating radar signal to the scanning position to form a data set.
S2, simulating a model test of grouting behind a shield tunnel wall of the ground penetrating radar, acquiring a ground penetrating radar image of the shield tunnel with known grouting thickness, and constructing a data set, wherein the method specifically comprises the following steps of:
s21, selecting shield tunnel segments and grouting slurry for model test, and according to the distribution 1 of the segments and grouting: 1, making a model, placing a duct piece in the same soil, injecting the same slurry, and placing a wave-absorbing material under boundary condition treatment;
s22, the thickness of the grouting layer in the model is divided into three types of undergrouting, normal grouting and over grouting according to engineering ground conditions and engineering experience;
s23, performing a model test by using the ground penetrating radar, and scanning grouting layers with known thickness classifications;
s24, recording the scanning position and scanning path of the radar, collecting signals of the ground penetrating radar, marking the corresponding position of each signal on the scanning path, and dividing the signals into three types of undergrouting, normal grouting and over grouting.
S3, after preprocessing data, according to the characteristics of a data set, establishing a prediction model of XGboost grouting thickness based on a ground penetrating radar image, as shown in FIG. 2, specifically comprising the following steps:
s31, carrying out data preprocessing on radar signals acquired in a model test and a shield tunnel, wherein the data preprocessing comprises the steps of drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization in sequence (the obtained radar signals are data between [ -1,1] and form a sample set;
s32, dividing a sample set into a training sample set I, a verification sample set and a test sample set according to a random mode for radar data collected by a preprocessed model test;
s33, calculating a loss function of the training sample set I in the current round according to the training sample set I:
Figure 863074DEST_PATH_IMAGE058
；
in the method, in the process of the invention,
Figure 469505DEST_PATH_IMAGE059
for loss function->
Figure 555273DEST_PATH_IMAGE060
For training sample number, ++>
Figure 438915DEST_PATH_IMAGE061
For->
Figure 270211DEST_PATH_IMAGE062
Training samples
Figure 239304DEST_PATH_IMAGE063
Loss of->
Figure 863184DEST_PATH_IMAGE064
For training the true tag value of the sample, +.>
Figure 991546DEST_PATH_IMAGE065
Is the firstiThe first sample is att-1Strong learner prediction value at multiple iterations,/->
Figure 636154DEST_PATH_IMAGE066
For the ith training sample->
Figure 702330DEST_PATH_IMAGE067
In the first placetA weak learner function during the training of the second iteration,/->
Figure 457796DEST_PATH_IMAGE068
And->
Figure 473288DEST_PATH_IMAGE069
For manually set coefficients +.>
Figure 164163DEST_PATH_IMAGE070
For the number of leaf nodes, +.>
Figure 842269DEST_PATH_IMAGE071
Is->
Figure 526060DEST_PATH_IMAGE072
A leaf node value;
s34, calculating an experience loss function of the current sample according to the training sample set I
Figure 910905DEST_PATH_IMAGE073
First and second partial derivatives based on a previous machine learning, and summing the first and second partial derivatives:
Figure 631737DEST_PATH_IMAGE074
；
in the method, in the process of the invention,
Figure 420308DEST_PATH_IMAGE075
for loss function->
Figure 189681DEST_PATH_IMAGE076
To fall into->
Figure 22507DEST_PATH_IMAGE077
Training sample set of individual leaf nodes, +.>
Figure 773295DEST_PATH_IMAGE078
Is->
Figure 691572DEST_PATH_IMAGE079
Loss function pair of individual samples +.>
Figure 264636DEST_PATH_IMAGE079
First partial derivative of the predicted value of the individual samples, ->
Figure 312488DEST_PATH_IMAGE080
Is->
Figure 375122DEST_PATH_IMAGE079
Loss function pair of individual samples +.>
Figure 390483DEST_PATH_IMAGE079
Second partial derivative of the predicted value of the individual samples,>
Figure 95134DEST_PATH_IMAGE081
and->
Figure 558345DEST_PATH_IMAGE082
For manually set coefficients, empirical loss functionsLIs part of the loss function, i.e., the remainder of the regularization is removed;
s35, the number of passing wheels istIteration of =1, 2..t, establish XGboost predictionA model comprising the steps of:
s351, calculate the first
Figure 667246DEST_PATH_IMAGE083
The individual samples are in the empirical loss function->
Figure 294537DEST_PATH_IMAGE084
Based on->
Figure 957206DEST_PATH_IMAGE085
Is>
Figure 963339DEST_PATH_IMAGE086
And second derivative->
Figure 633355DEST_PATH_IMAGE087
Calculate the sum of the first derivatives of all samples +.>
Figure 669313DEST_PATH_IMAGE088
Sum of second derivatives of all samples +.>
Figure 59974DEST_PATH_IMAGE089
，=1，2，…，
Figure 999242DEST_PATH_IMAGE090
；
S352, based on the current node attempting to split the decision tree, for all featureskCalculating the maximum score, which specifically comprises the following steps:
s3521 making the initial maximum score
Figure 574580DEST_PATH_IMAGE091
，
Figure 848567DEST_PATH_IMAGE092
、
Figure 88924DEST_PATH_IMAGE093
；
S3522 sample is characterized by
Figure 69650DEST_PATH_IMAGE042
Sequentially taking out the +.>
Figure 513093DEST_PATH_IMAGE094
And sequentially calculating the sum of the first derivatives and the second derivatives of the left subtree and the right subtree after the current sample is placed in the left subtree:
Figure 71114DEST_PATH_IMAGE095
，
Figure 131474DEST_PATH_IMAGE096
；
Figure 215973DEST_PATH_IMAGE097
，
Figure 70797DEST_PATH_IMAGE098
；
in the method, in the process of the invention,
Figure 647271DEST_PATH_IMAGE099
representing a numerical update->
Figure 730896DEST_PATH_IMAGE100
Sum of all sample first derivatives of left subtree,/->
Figure 420635DEST_PATH_IMAGE101
Sum of all samples first derivatives of right subtree, +.>
Figure 226786DEST_PATH_IMAGE102
For the sum of the first derivatives of all samples, +.>
Figure 900343DEST_PATH_IMAGE103
Sum of all sample first derivatives of left subtree,/->
Figure 50309DEST_PATH_IMAGE104
Is the sum of the second derivatives of all samples of the right subtree, < ->
Figure 719187DEST_PATH_IMAGE105
Sum of second derivatives for all samples;
Figure 181393DEST_PATH_IMAGE106
And->
Figure 529198DEST_PATH_IMAGE107
The first derivative and the second derivative of the ith sample of the t-th round entering the left subtree are respectively;
s3523, updating the maximum score:
Figure 531789DEST_PATH_IMAGE108
；
s353 based on maximum
Figure 9169DEST_PATH_IMAGE109
Splitting subtrees by corresponding dividing features and feature values;
s354, if maximum
Figure 845538DEST_PATH_IMAGE109
If the value is 0, the current decision tree is built, and the +.>
Figure 87163DEST_PATH_IMAGE110
，
Figure 80396DEST_PATH_IMAGE110
Is the firsttWheel training firstjThe values of the individual leaf nodes, obtaining weak learner->
Figure 864812DEST_PATH_IMAGE111
Updating strong learner->
Figure 151044DEST_PATH_IMAGE112
Enter underA round of weak learner iteration if max +.>
Figure 145544DEST_PATH_IMAGE113
If not 0, then the process proceeds to step S352 to continue with the attempt to split the decision tree.
S4, optimizing parameters of the prediction model of the XGboost grouting thickness, as shown in fig. 3 and 4, specifically comprises the following steps:
s41, introducing K-fold cross verification when the XGBoost algorithm is trained, estimating a generalization error of a classifier, and evaluating the classification accuracy of the XGBoost algorithm under different parameters when only a training sample is possessed;
s42, searching the multidimensional array in parallel from different growth directions by adopting a grid search algorithm, and determining parameters of grid search, wherein the method comprises the following steps of:
s421, determining a search range of parameters to be searched according to experience;
s422, setting a search step length of the searched parameter;
s423, calculating the classification accuracy of XGBoost according to a cross verification method for each group of values on the grid;
and S424, giving the classification accuracy of each group, and determining the optimal parameter value according to the classification accuracy.
S5, predicting a ground penetrating radar image acquired by the grouting ground penetrating radar in real time after the shield tunnel wall by using a prediction model of the XGboost grouting thickness after parameter optimization, and specifically comprising the following steps:
s51, preprocessing the ground penetrating radar data acquired in the S1, wherein the steps of data preprocessing sequentially comprise drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization, and the obtained radar signals are data between [ -1,1 ];
s52, inputting the data into an established XGboost prediction model to obtain a prediction value of the grouting layer thickness classification, namely outputting a classification label finally in FIG. 1.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, and various modifications and variations may be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. A prediction method of post grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning comprises the following steps:
s1, acquiring signals and images in a shield tunnel by using a ground penetrating radar, and corresponding the ground penetrating radar signals to scanning positions to form a data set;
s2, simulating a model test of grouting behind a shield tunnel wall of the ground penetrating radar, acquiring a ground penetrating radar image of the shield tunnel with known grouting thickness, and constructing a data set;
s3, after preprocessing the data, establishing a prediction model of XGboost grouting thickness based on the ground penetrating radar image according to the characteristics of the data set;
the step S3 includes the steps of:
s31, carrying out data preprocessing on radar signals acquired in a model test and a shield tunnel, wherein the data preprocessing comprises the steps of drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization in sequence, and the obtained radar signals are data between [ -1,1] and form a sample set;
s32, dividing a sample set into a training sample set I, a verification sample set and a test sample set according to a random mode for radar data collected by a preprocessed model test;
s33, calculating a loss function of the training sample set I in the current round according to the training sample set I:
Figure DEST_PATH_IMAGE001
；
in the method, in the process of the invention,
Figure 427836DEST_PATH_IMAGE002
for loss function->
Figure DEST_PATH_IMAGE003
For training sample number, ++>
Figure 157895DEST_PATH_IMAGE004
For->
Figure DEST_PATH_IMAGE005
Training samples->
Figure 561194DEST_PATH_IMAGE006
Loss of->
Figure DEST_PATH_IMAGE007
For training the true tag value of the sample, +.>
Figure 99098DEST_PATH_IMAGE008
Is->
Figure 587848DEST_PATH_IMAGE005
The first sample is att-1Strong learner prediction value at multiple iterations,/->
Figure DEST_PATH_IMAGE009
Is->
Figure 856019DEST_PATH_IMAGE005
Training samples->
Figure 176142DEST_PATH_IMAGE006
In the first placetA weak learner function during the training of the second iteration,/->
Figure 605986DEST_PATH_IMAGE010
And
Figure DEST_PATH_IMAGE011
for manually set coefficients +.>
Figure 378770DEST_PATH_IMAGE012
For the number of leaf nodes, +.>
Figure DEST_PATH_IMAGE013
Is->
Figure 388314DEST_PATH_IMAGE014
A leaf node value;
s34, calculating an experience loss function of the current sample according to the training sample set I
Figure DEST_PATH_IMAGE015
First and second partial derivatives based on a previous machine learning, and summing the first and second partial derivatives:
Figure 579255DEST_PATH_IMAGE016
；
in the method, in the process of the invention,
Figure DEST_PATH_IMAGE017
for loss function->
Figure 445580DEST_PATH_IMAGE018
To fall into->
Figure DEST_PATH_IMAGE019
Training sample set of individual leaf nodes, +.>
Figure 908922DEST_PATH_IMAGE020
Is->
Figure DEST_PATH_IMAGE021
Loss function pair of individual samples +.>
Figure 518895DEST_PATH_IMAGE021
First partial derivative of the predicted value of the individual samples, ->
Figure 16873DEST_PATH_IMAGE022
Is->
Figure 585257DEST_PATH_IMAGE021
Loss function pair of individual samples +.>
Figure 270317DEST_PATH_IMAGE021
Second partial derivative of the predicted value of the individual samples,>
Figure DEST_PATH_IMAGE023
and->
Figure 700292DEST_PATH_IMAGE024
For manually set coefficients, empirical loss functionsLIs part of the loss function, i.e., the remainder of the regularization is removed;
s35, the number of passing wheels istIteration of =1, 2..t, build XGboost prediction model;
s4, optimizing parameters of the prediction model of the XGboost grouting thickness;
the step S4 includes the steps of:
s41, introducing K-fold cross verification when the XGBoost algorithm is trained, estimating a generalization error of a classifier, and evaluating the classification accuracy of the XGBoost algorithm under different parameters when only a training sample is possessed;
s42, searching the multidimensional array in parallel from different growth directions by adopting a grid search algorithm, and determining parameters of grid search;
the step S42 includes the steps of:
s421, determining a search range of parameters to be searched according to experience;
s422, setting a search step length of the searched parameter;
s423, calculating the classification accuracy of XGBoost according to a cross verification method for each group of values on the grid;
s424, giving the classification accuracy of each group, and determining the optimal parameter value according to the classification accuracy;
s5, predicting a ground penetrating radar image acquired by the grouting ground penetrating radar in real time after the shield tunnel wall by using a prediction model of the XGboost grouting thickness after parameter optimization.
2. The prediction method of the post-grouting thickness of the shield tunnel wall based on ground penetrating radar detection and machine learning according to claim 1, wherein the prediction method is characterized by comprising the following steps: the step S1 includes the steps of:
s11, using a ground penetrating radar to scan and detect lining of the tail of a shield tunnel and grouting behind the wall of the shield tunnel under construction, wherein the ground penetrating radar is positioned close to the duct piece;
s12, recording the scanning position and the scanning path of the ground penetrating radar, collecting signals of the ground penetrating radar, and marking the corresponding position of each signal on the scanning path;
s13, numbering each acquired radar signal from 0;
and S14, corresponding the ground penetrating radar signal to the scanning position to form a data set.
3. The prediction method of the post-grouting thickness of the shield tunnel wall based on ground penetrating radar detection and machine learning according to claim 1, wherein the prediction method is characterized by comprising the following steps: the step S2 includes the steps of:
s21, selecting shield tunnel segments and grouting slurry for model test, and according to the distribution 1 of the segments and grouting: 1, making a model, placing a duct piece in the same soil, injecting the same slurry, and placing a wave-absorbing material under boundary condition treatment;
s22, the thickness of the grouting layer in the model is divided into three types of undergrouting, normal grouting and over grouting according to engineering ground conditions and engineering experience;
s23, performing a model test by using the ground penetrating radar, and scanning grouting layers with known thickness classifications;
s24, recording the scanning position and scanning path of the radar, collecting signals of the ground penetrating radar, marking the corresponding position of each signal on the scanning path, and dividing the signals into three types of undergrouting, normal grouting and over grouting.
4. The prediction method of the post-grouting thickness of the shield tunnel wall based on ground penetrating radar detection and machine learning according to claim 1, wherein the prediction method is characterized by comprising the following steps: the step S35 includes the steps of:
s351, calculate the first
Figure 52776DEST_PATH_IMAGE026
The individual samples are in the empirical loss function->
Figure 729745DEST_PATH_IMAGE028
Based on->
Figure 964417DEST_PATH_IMAGE030
Is>
Figure 385034DEST_PATH_IMAGE032
And second derivative->
Figure 592025DEST_PATH_IMAGE034
Calculate the sum of the first derivatives of all samples +.>
Figure 767791DEST_PATH_IMAGE036
Sum of second derivatives of all samples +.>
Figure 240492DEST_PATH_IMAGE038
，
Figure 199221DEST_PATH_IMAGE040
，
Figure 260718DEST_PATH_IMAGE042
；
S352, based on the current node attempting to split the decision tree, for all featureskCalculating a maximum score;
s353 based on maximum
Figure 341806DEST_PATH_IMAGE044
Splitting subtrees by corresponding dividing features and feature values;
s354, if maximum
Figure 488754DEST_PATH_IMAGE044
If the value is 0, the current decision tree is built, and the +.>
Figure 516752DEST_PATH_IMAGE046
，
Figure 495073DEST_PATH_IMAGE046
Is the firsttWheel training firstjThe values of the individual leaf nodes, obtaining weak learner->
Figure 419166DEST_PATH_IMAGE048
Updating strong learner->
Figure DEST_PATH_IMAGE050
Enter the next round of weak learner iteration if max +>
Figure 115727DEST_PATH_IMAGE044
If not 0, then the process proceeds to step S352 to continue with the attempt to split the decision tree. />
5. The prediction method of the post-grouting thickness of the shield tunnel wall based on ground penetrating radar detection and machine learning according to claim 4, wherein the prediction method is characterized by comprising the following steps: the step S352 includes the steps of:
s3521 making the initial maximum score
Figure DEST_PATH_IMAGE052
，
Figure DEST_PATH_IMAGE054
、
Figure DEST_PATH_IMAGE056
；
S3522 sample is characterized by
Figure DEST_PATH_IMAGE058
Sequentially taking out the +.>
Figure DEST_PATH_IMAGE060
And sequentially calculating the sum of the first derivatives and the second derivatives of the left subtree and the right subtree after the current sample is placed in the left subtree:
Figure DEST_PATH_IMAGE062
，
Figure DEST_PATH_IMAGE064
；
Figure DEST_PATH_IMAGE066
，
Figure DEST_PATH_IMAGE068
；
in the method, in the process of the invention,
Figure DEST_PATH_IMAGE070
representing a numerical update->
Figure DEST_PATH_IMAGE072
Sum of all sample first derivatives of left subtree,/->
Figure DEST_PATH_IMAGE074
Sum of all samples first derivatives of right subtree, +.>
Figure DEST_PATH_IMAGE076
For the sum of the first derivatives of all samples，
Figure DEST_PATH_IMAGE078
Sum of all sample first derivatives of left subtree,/->
Figure DEST_PATH_IMAGE080
Is the sum of the second derivatives of all samples of the right subtree, < ->
Figure DEST_PATH_IMAGE082
Sum of second derivatives for all samples;
Figure DEST_PATH_IMAGE084
And
Figure DEST_PATH_IMAGE086
the first derivative and the second derivative of the ith sample of the t-th round entering the left subtree are respectively;
s3523, updating the maximum score:
Figure DEST_PATH_IMAGE088
。
6. the prediction method of the post-grouting thickness of the shield tunnel wall based on ground penetrating radar detection and machine learning according to claim 1, wherein the prediction method is characterized by comprising the following steps: the step S5 includes the steps of:
s51, preprocessing ground penetrating radar data acquired in the S1, wherein the data preprocessing step comprises drift removal, butterworth band-pass filtering, moving average, F-K offset and normalization, and the obtained radar signals are data between [ -1,1 ];
s52, inputting the data into an established XGboost prediction model to obtain a prediction value of the grouting layer thickness classification.
CN202010975476.2A 2020-09-16 2020-09-16 Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning Active CN112287595B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010975476.2A CN112287595B (en) 2020-09-16 2020-09-16 Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010975476.2A CN112287595B (en) 2020-09-16 2020-09-16 Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning

Publications (2)

Publication Number Publication Date
CN112287595A CN112287595A (en) 2021-01-29
CN112287595B true CN112287595B (en) 2023-05-23

Family

ID=74419965

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010975476.2A Active CN112287595B (en) 2020-09-16 2020-09-16 Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning

Country Status (1)

Country Link
CN (1) CN112287595B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113326661B (en) * 2021-06-17 2022-02-15 广西路桥工程集团有限公司 RS-XGboost-based tunnel advanced drilling quantitative interpretation method and device
CN113326660B (en) * 2021-06-17 2022-11-29 广西路桥工程集团有限公司 Tunnel surrounding rock extrusion deformation prediction method based on GA-XGboost model

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1975112A (en) * 2006-12-14 2007-06-06 同济大学 Shield tunnel subsidence control method based on exploring radar
CN111025286A (en) * 2019-11-01 2020-04-17 长安大学 Ground penetrating radar map self-adaptive selection method for water damage detection
CN111144636A (en) * 2019-12-22 2020-05-12 同济大学 Slope deformation prediction method
CN111365015A (en) * 2020-03-05 2020-07-03 中建交通建设集团有限公司 Shield tunneling parameter feature extraction and attitude deviation prediction method based on XGboost

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1975112A (en) * 2006-12-14 2007-06-06 同济大学 Shield tunnel subsidence control method based on exploring radar
CN111025286A (en) * 2019-11-01 2020-04-17 长安大学 Ground penetrating radar map self-adaptive selection method for water damage detection
CN111144636A (en) * 2019-12-22 2020-05-12 同济大学 Slope deformation prediction method
CN111365015A (en) * 2020-03-05 2020-07-03 中建交通建设集团有限公司 Shield tunneling parameter feature extraction and attitude deviation prediction method based on XGboost

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"Real-time detection technology of synchronous grouting for shield tunnel and analysis of grouting effect";Xie Xiongyao el.;《2018 17th International Conference on Ground Penetrating Radar (GPR)》;20180823;全文 *
"基于机器学习的地铁隧道施工扰动控制研究";王鹏;《现代隧道技术》;20191231;第56卷;全文 *

Also Published As

Publication number Publication date
CN112287595A (en) 2021-01-29

Similar Documents

Publication Publication Date Title
Ukhwah et al. Asphalt pavement pothole detection using deep learning method based on YOLO neural network
KR101811270B1 (en) Method and system for checking goods
CN114548278A (en) In-service tunnel lining structure defect identification method and system based on deep learning
CN112287595B (en) Prediction method of post-grouting thickness of shield tunnel wall based on ground penetrating radar detection and machine learning
CN115019171B (en) Non-contact surrounding rock rapid partition grading method
CN112446350B (en) Improved method for detecting cotton in YOLOv3 complex cotton field background
CN111210869A (en) Protein cryoelectron microscope structure analysis model training method and analysis method
Jiang et al. Development of a core feature identification application based on the Faster R-CNN algorithm
CN117741808A (en) Geological crack detection method based on geological exploration
CN116468939A (en) Intelligent classification method for tunnel face surrounding rock based on neural network
Fakhri et al. Road crack detection using gaussian/prewitt filter
CN111783616A (en) Data-driven self-learning-based nondestructive testing method
CN113095364B (en) High-speed rail seismic event extraction method, medium and equipment using convolutional neural network
CN107437091A (en) Multilayer limits the positive and negative class change detecting method of SAR image of Boltzmann machine
CN117349462B (en) Remote sensing intelligent interpretation sample data set generation method
CN108663334A (en) The method for finding soil nutrient spectral signature wavelength based on multiple Classifiers Combination
CN116466408B (en) Artificial neural network superbedrock identification method based on aeromagnetic data
Valencia et al. Blasthole Location Detection Using Support Vector Machine and Convolutional Neural Networks on UAV Images and Photogrammetry Models
CN117688313A (en) Method for cooperatively processing mapping data by network
CN117496463A (en) Learning method and system for improving road perception precision
CN115860396A (en) Intelligent planning method based on land reserve analysis model
CN114724036A (en) Ice crack detection method based on multi-scale learning
Vanitha et al. A novel deep learning method for the identification and categorization of footpath defects based on thermography
CN116597296B (en) Wheat seedling density estimation method based on deep learning and thermodynamic diagram
CN113963334B (en) Vehicle track defect data repairing method for road side sensing unit

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant