CN114996829A - Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel - Google Patents

Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel Download PDF

Info

Publication number
CN114996829A
CN114996829A CN202210917562.7A CN202210917562A CN114996829A CN 114996829 A CN114996829 A CN 114996829A CN 202210917562 A CN202210917562 A CN 202210917562A CN 114996829 A CN114996829 A CN 114996829A
Authority
CN
China
Prior art keywords
tunnel
model
value
pareto
newly
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.)
Withdrawn
Application number
CN202210917562.7A
Other languages
Chinese (zh)
Inventor
张立茂
吴贤国
刘琼
林鹏辉
覃亚伟
徐文胜
张军
姚春桥
王金峰
曾铁梅
陶文涛
熊朝辉
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.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
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 Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202210917562.7A priority Critical patent/CN114996829A/en
Publication of CN114996829A publication Critical patent/CN114996829A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • 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
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Geometry (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computer Hardware Design (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Civil Engineering (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Genetics & Genomics (AREA)
  • Physiology (AREA)
  • Architecture (AREA)
  • Biomedical Technology (AREA)
  • Structural Engineering (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention belongs to the technical field of tunnel construction, and particularly discloses a newly-built tunnel design optimization method and equipment under the condition of near tunnel construction. The method comprises the following steps: constructing a tunnel construction parameter index system, and establishing an integrated learning model for predicting ultimate support pressure and earth surface deformation by adopting a lightweight gradient promoting machine algorithm; optimizing two targets of limit supporting pressure and earth surface deformation by using NSGA-II, and simultaneously considering uncertainty of rock and soil conditions and errors of the meta-model; and establishing probability constraint to perform multi-objective optimization through Monte Carlo simulation to generate a Pareto front edge, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion. The method can predict and optimize the excavation damage of the adjacent existing tunnel, the optimal position of the new tunnel is determined on the basis of the existing tunnel, and various uncertainties such as geological conditions, meta-models and the like are considered, so that the result is more reliable, and more conservative reference is provided for decision making.

Description

Method and equipment for optimizing design of newly-built tunnel under construction condition of close-proximity tunnel
Technical Field
The invention belongs to the technical field of tunnel construction, and particularly relates to a method and equipment for optimizing design of a newly-built tunnel under a close-proximity tunnel construction condition.
Background
In tunnel construction problems, it is often necessary to construct adjacent existing tunnels, which may involve two tunnels running across each other or running in parallel. In this case, it is necessary to consider the existence of the adjacent tunnel at the initial stage of the design to find a suitable position for the construction of the new tunnel. Due to the rapid development of ensemble learning algorithms in recent years, such as Random Forest (RF), gradient enhanced decision tree (GBDT), extreme gradient enhancement (LightGBM), existing data sets can now be used to predict tunnel excavation damage. Aiming at the problem of tunnel construction, the machine learning method can fully and reasonably simulate the damage caused by construction, but few methods can relieve the damage of tunnel construction. The goal of multi-objective optimization (MOO) is to achieve a balance between all objectives, so as to find the best solution. The modified version of NSGA-II is one of the most popular MOO methods, and is enhanced in computational speed and population elite performance. However, such algorithms typically use deterministic constraints, lacking consideration of the uncertainty of the variables. The uncertainty of the soil comes from one specific location (local uncertainty) or multiple locations (spatial uncertainty. if the differences in soil properties are not properly accounted for, the reaction may exceed the design's capability, causing a huge loss.
Based on the defects and shortcomings, the field needs to provide a newly-built tunnel design optimization method under the construction condition of the close-coupled tunnel, consider the uncertainty of geotechnical conditions, obtain the optimal position of the new tunnel by taking the point with the shortest distance from an ideal point as a criterion, and simultaneously predict and optimize the excavation damage of the adjacent existing tunnel.
Disclosure of Invention
Aiming at the defects or improvement requirements of the prior art, the invention provides a newly-built tunnel design optimization method and equipment under the construction condition of a proximity tunnel, wherein an integrated learning model for predicting ultimate support pressure (LSP) and Ground Surface Deformation (GSD) is established by constructing a tunnel construction parameter index system and adopting a lightweight gradient booster (LightGBM) algorithm; optimizing two targets of LSP and GSD by using NSGA-II, and simultaneously considering uncertainty of rock and soil conditions and errors of the meta-model; and establishing probability constraint to perform multi-objective optimization (MOO) through Monte Carlo simulation, generating Pareto leading edges, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion. Excavation damage of adjacent existing tunnels can be predicted and optimized, and various types of uncertainty such as geological conditions and meta-models are considered, so that reliability is guaranteed.
In order to achieve the above object, according to an aspect of the present invention, a method for optimizing a newly-built tunnel design under a near tunnel construction condition is provided, which includes the following steps:
s1, constructing a tunnel construction parameter index system, and establishing an integrated learning model for predicting ultimate support pressure and earth surface deformation by adopting a LightGBM regression model based on the tunnel construction parameter index system;
s2, optimizing limit supporting pressure and earth surface deformation by using uncertainty of rock and soil conditions and meta-model errors of the integrated learning model as constraints;
s3, probability constraint is established, multi-objective optimization is carried out on the integrated learning model, a Pareto front edge is generated, an optimal solution is selected by taking the point with the shortest ideal point distance as a criterion, and the optimal solution is the optimal position of the new tunnel.
More preferably, step S1 specifically includes: the method comprises the steps of collecting an existing data set, dividing the existing data set into a training set and a testing set, adopting the training set to carry out prediction precision analysis on a LightGBM regression model, and adopting mean square error and root mean square error to evaluate the accuracy of the LightGBM regression model prediction.
More preferably, in step S2, the data of the existing data set is used as y, and the predicted value of the LightGBM regression model is used as y
Figure 514300DEST_PATH_IMAGE001
Training a linear regression model by using a least square method as follows:
Figure 585025DEST_PATH_IMAGE002
in the formula,
Figure 638431DEST_PATH_IMAGE003
substitution for regression equationsxThe estimated value obtained is used as a reference value,
Figure 884736DEST_PATH_IMAGE004
in order to be a function of the regression equation,
Figure 896554DEST_PATH_IMAGE005
for existing data setiThe number of the data is one,
Figure 720154DEST_PATH_IMAGE006
to predict the valueiThe number of the data is set to be,
Figure 311672DEST_PATH_IMAGE007
is a predicted value of the regression model,
Figure 661751DEST_PATH_IMAGE008
for the average of the existing data set,iis a sample number.
Preferably, the LightGBM regression model, wherein the quantifying the uncertainty of the meta-model using the prediction interval of the regression process, comprises: in the linear regression model, a given sample volume isnAs a sample of an existing data setyPredicted value of LightGBM regression model asxCorresponding prediction intervalPICalculated from the following formula:
Figure 578891DEST_PATH_IMAGE009
wherein,
Figure 624208DEST_PATH_IMAGE010
is based on
Figure 19417DEST_PATH_IMAGE011
The predicted value of the regression model is determined,
Figure 99369DEST_PATH_IMAGE012
is composed of
Figure 62776DEST_PATH_IMAGE013
Degree of freedom
Figure 595389DEST_PATH_IMAGE014
At a significant leveltThe value of the one or more of the one,
Figure 59868DEST_PATH_IMAGE015
in order to be the standard error of the residual error,
Figure 728747DEST_PATH_IMAGE016
and
Figure 987690DEST_PATH_IMAGE017
are respectively asxThe mean and the variance of (a) is,nis the sample volume.
As a further preferable method, in step S2, the method for quantifying uncertainty from geotechnical conditions using monte carlo simulation method includes: assuming that the soil parameters under consideration obey a certain probability distribution, it will resultnThe samples are subjected to the distribution, at a prediction input variable of
Figure 132232DEST_PATH_IMAGE018
The Monte Carlo simulation method calculates the output value with each sample, all results are within a defined range, expressed asmThus predicting a probability within a defined range of
Figure 400403DEST_PATH_IMAGE019
。
Preferably, in step S3, establishing a probability constraint to perform a multi-objective optimization on the ensemble learning model, the method includes:
Figure 923788DEST_PATH_IMAGE020
wherein,
Figure 353632DEST_PATH_IMAGE021
is the firstjA function of a constraint that is a function of the constraint,
Figure 860837DEST_PATH_IMAGE022
is a defined probability limit for the probability of,
Figure 401540DEST_PATH_IMAGE023
in order to integrate the learning model function,
Figure 920377DEST_PATH_IMAGE024
is as followsmAn objective function of the number of the target functions,
Figure 786702DEST_PATH_IMAGE025
as the first in the data setiThe number of the input variables is changed according to the number of the input variables,
Figure 515623DEST_PATH_IMAGE026
is as followsiThe minimum value of the number of the input variables,
Figure 125596DEST_PATH_IMAGE027
is as followsiMaximum value of each input variable.
As a further preferred, in step S3, calculating Pareto frontier of the multi-objective optimization problem using NSAG-ii by defining probability constraints considering soil properties and meta-model uncertainty in geotechnical conditions, includes: is shown astThe parent population of the generations is
Figure 357994DEST_PATH_IMAGE028
Through selection, crossover and mutation, progeny populations will be generated
Figure 519854DEST_PATH_IMAGE029
Suppose that
Figure 736072DEST_PATH_IMAGE030
And
Figure 618577DEST_PATH_IMAGE031
all the members are N, the parent population and the offspring population are combined to carry out sequencing, the elite property of NSGA-II is ensured, and all the members are firstly sorted according to the formulaNon-dominant rank ordering and selecting from lowest ordering to next generation untilNThe positions are filled, wherein the members belonging to the previous stage are further sorted according to the congestion distance, preferably passed on to the next generation, the process is repeated until the last generation, and the Pareto frontier is selected as
Figure 236641DEST_PATH_IMAGE032
；
According to a group of optimal solutions provided by the obtained Pareto frontier, sorting the Pareto optimal solutions, including: selecting the optimal point with the shortest distance to the ideal point, and solving the problem of multi-objective optimizationiThe ideal value of each target is represented as
Figure 444768DEST_PATH_IMAGE033
Then put pareto aheadjThe distance of the solution to the ideal point is defined as:
Figure 23648DEST_PATH_IMAGE034
wherein,mis a target total number of the cells,
Figure 975424DEST_PATH_IMAGE035
is shown asjSecond in Pareto solution
Figure 182414DEST_PATH_IMAGE036
The value of the individual object is,
Figure 561443DEST_PATH_IMAGE037
is as followsiThe desired value of the individual target.
According to another aspect of the present invention, there is also provided a system for optimizing a design of a newly constructed tunnel under a condition of a close-proximity tunnel construction, comprising: the first main module is used for constructing a shield construction parameter index system and acquiring the existing data; the second main module is used for constructing and training a LightGBM regression model, and establishing an integrated learning model for predicting ultimate support pressure and surface deformation by adopting a LightGBM algorithm according to the existing data set; the third main module is used for optimizing two targets of limit supporting pressure and earth surface deformation by using NSGA-II, and simultaneously using uncertainty of rock-soil conditions and meta-model errors of the integrated learning model as constraints; and the fourth main module is used for establishing probability constraint through a Monte Carlo simulation method, performing multi-objective optimization on the integrated learning model, generating a Pareto front edge, and obtaining the optimal position of the new tunnel by taking a point with the shortest distance from an ideal point as a criterion.
According to another aspect of the present invention, there is also provided an electronic apparatus including:
at least one processor, at least one memory, and a communication interface; wherein,
the processor, the memory and the communication interface are communicated with each other;
the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute a newly-built tunnel design optimization method under the condition of close tunnel construction.
According to another aspect of the present invention, there is also provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method for optimizing a design of a newly-built tunnel under a proximate tunnel construction condition.
Generally, compared with the prior art, the above technical solution conceived by the present invention mainly has the following technical advantages:
1. the method adopts a lightweight gradient booster (LightGBM) algorithm to establish an integrated learning model for predicting ultimate support pressure (LSP) and surface deformation (GSD); optimizing two targets of LSP and GSD, and simultaneously considering uncertainty of rock-soil conditions and errors of the meta-model; and establishing probability constraint to carry out multi-objective optimization (MOO), generating a Pareto front edge, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion. Excavation damage of adjacent existing tunnels can be predicted and optimized, and various types of uncertainty such as geological conditions and meta-models are considered, so that reliability is guaranteed.
2. The invention develops a mixed method combining LightGBM and NSGA-II to predict and optimize the damage caused by the existing tunnel excavation, and adopts a Monte Carlo simulation method to simulate the uncertainty of the soil property. Finally, the optimal position of the new tunnel will be determined on the basis of the existing tunnels. The invention considers the uncertainty of the soil and the meta-model, so that the result is more reliable, and a more conservative reference is provided for decision making.
Drawings
Fig. 1 is a flowchart of a design optimization method for a newly built tunnel under a proximate tunnel construction condition according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a newly-built tunnel design optimization device under a proximate tunnel construction condition according to an embodiment of the present invention;
fig. 3 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention;
fig. 4 shows 10 factor characteristics of a tunnel excavation project according to an embodiment of the present invention;
fig. 5 is MAE and RMSE of training data and test data corresponding to LSP and GSD provided in the embodiment of the present invention;
fig. 6 is a Pareto front schematic diagram of an MOO under a specific condition without considering uncertainty when excavation lengths provided by the embodiment of the present invention are 10m, 35m, and 60m, respectively;
fig. 7 is a schematic diagram of the resulting position of the Pareto front under a specific condition without considering uncertainty when the excavation lengths provided by the embodiment of the present invention are 10m, 35m, and 60m, respectively;
fig. 8 is a Pareto front diagram of an MOO under a specific condition considering uncertainty (95% PI) when excavation lengths provided by an embodiment of the present invention are 10m, 35m, and 60m, respectively;
fig. 9 is a schematic diagram of the resulting position of the Pareto front in a specific case considering uncertainty (95% PI) when the excavation lengths provided by the embodiment of the present invention are 10m, 35m, and 60m, respectively;
fig. 10 is a schematic diagram of the percentage improvement of 20 test cases in different scenarios of LSP and GSD without considering uncertainty and with considering uncertainty (95% PI) according to the embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention. In addition, technical features of various embodiments or individual embodiments provided by the present invention may be arbitrarily combined with each other to form a feasible technical solution, and such combination is not limited by the sequence of steps and/or the structural composition mode, but must be realized by a person skilled in the art, and when the technical solution combination is contradictory or cannot be realized, such a technical solution combination should not be considered to exist and is not within the protection scope of the present invention.
The embodiment of the invention provides a newly-built tunnel design optimization method under the condition of close-proximity tunnel construction, and with reference to a figure 1, the method comprises the following steps: constructing a tunnel construction parameter index system, and establishing an integrated learning model for predicting ultimate support pressure and earth surface deformation by adopting a LightGBM regression model based on the tunnel construction parameter index system; optimizing limit support pressure and earth surface deformation by using uncertainty of rock and soil conditions and meta-model errors of the integrated learning model as constraints; and establishing probability constraint, performing multi-objective optimization on the integrated learning model, generating a Pareto front edge, and selecting an optimal solution by taking the point with the shortest ideal point distance as a criterion, wherein the optimal solution is the optimal position of the new tunnel. More specifically, in the method, a tunnel construction parameter index system is firstly established, and an integrated learning model for predicting ultimate support pressure (LSP) and Ground Surface Deformation (GSD) is established by adopting a lightweight gradient booster (LightGBM) algorithm according to an existing data set; optimizing two targets of LSP and GSD by using NSGA-II, and simultaneously considering uncertainty of rock and soil conditions and errors of the meta-model; and establishing probability constraint to perform multi-objective optimization (MOO) through Monte Carlo simulation, generating Pareto leading edges, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion.
Based on the content of the foregoing method embodiment, as an optional embodiment, the method for optimizing design of a newly-built tunnel under a condition of approaching tunnel construction provided in the embodiment of the present invention, where the building of an ensemble learning model for predicting LSP and GSD by using a LightGBM algorithm according to an existing data set includes: and dividing the data sample set, randomly selecting four fifths of samples as a training set, remaining one fifths of samples as a test set, carrying out prediction precision analysis on the LightGBM regression model, and evaluating the prediction accuracy of the LightGBM regression model by adopting Mean Square Error (MSE) and Root Mean Square Error (RMSE).
Specifically, in order to simulate the underground environment of the adjacent tunnels, the uncertainty of the geotechnical conditions is considered, i.e., 10 features in two categories (soil properties and geometric relationships) are considered as input variables, and the features are summarized in fig. 4. However, in practical situations, other parameters than the location of the new tunnel have been determined. Therefore, even if a meta model including 10 factors is established, only the soil coverage (C) of the new tunnel and the horizontal distance (H) between the tunnels can be optimized. When the ensemble learning model is trained, 20 samples are randomly selected as test data. Of the remaining 80 samples, 10 were selected as validation sets, and the rest were modeled as training data sets. The hyper-parameters are then determined using a grid search cross-validation method. After the hyper-parameters are determined, the LightGBM model is trained to describe the LSP and GSD, an ensemble learning model is built and compared with other machine learning techniques, and the corresponding MAE and RMSE are shown in fig. 5. Compared to the multilayer perceptron (MLP) approach, LightGBM has a very high precision; the performance of Random Forest (RF) in the simulation of GSD is quite close to LightGBM, while the accuracy of LSP correspondence is low and the degree of model overfitting is high.
Based on the content of the method embodiment, as an optional embodiment, in the newly-built tunnel design optimization method under the condition of the close-proximity tunnel construction provided in the embodiment of the present invention, the uncertainty of the meta-model is quantified by using a Prediction Interval (PI) of the regression process. The method comprises the following steps: in a linear regression model describing y by x, given a sample of volume n, the corresponding Prediction Interval (PI) can be calculated by,
Figure 752253DEST_PATH_IMAGE038
wherein,
Figure 976561DEST_PATH_IMAGE039
is based on
Figure 693850DEST_PATH_IMAGE040
The predicted value of the regression model is determined,
Figure 978201DEST_PATH_IMAGE041
is composed of
Figure 656307DEST_PATH_IMAGE042
Degree of freedom
Figure 684305DEST_PATH_IMAGE014
The value of t at the significant level is,
Figure 131467DEST_PATH_IMAGE015
in order to be the standard error of the residual error,
Figure 727665DEST_PATH_IMAGE016
and
Figure 627488DEST_PATH_IMAGE043
are respectively asxThe mean and the variance of (a) is,nis the sample volume. As is clear from the above equation, a linear regression model must be established before the Prediction Interval (PI) is obtained. Taking the data of the original data set as y and the predicted value of the LightGBM regression model asxUsing the least squares (OLS) method, the following linear regression model can be trained:
Figure 193598DEST_PATH_IMAGE044
in the formula,
Figure 760846DEST_PATH_IMAGE045
an estimated value obtained by substituting x into the regression equation,
Figure 652578DEST_PATH_IMAGE046
in order to be a function of the regression equation,
Figure 774118DEST_PATH_IMAGE047
for the ith data of the existing data set,
Figure 534133DEST_PATH_IMAGE048
in order to predict the ith data of the value,
Figure 955887DEST_PATH_IMAGE049
is a predicted value of the regression model,
Figure 752941DEST_PATH_IMAGE050
for the existing data set mean, i is the sample number.
Specifically, to further test the performance of the meta-model, a linear regression model between the sample values and the predicted values is established:
Figure 627357DEST_PATH_IMAGE051
the Prediction Interval (PI) of the meta-model can be determined from the linear regression model. As baseline results, a significance level of 5% was used to obtain the interval. The application of the Prediction Interval (PI) can ensure the reliability of the machine learning model.
Based on the content of the method embodiment, as an optional embodiment, the newly-built tunnel design optimization method provided in the embodiment of the invention is used for quantifying uncertainty from geotechnical conditions by using a monte carlo simulation technology under the condition of proximate tunnel construction. The method comprises the following steps: assuming that the soil parameters under consideration obey a certain probability distribution, it will resultnThe samples obey this distribution. In predicting the input variable asxThe algorithm calculates the output value with each sample. All results are definedWithin the range, the defined range is indicated asmThus predicting a probability within a defined range of
Figure 800849DEST_PATH_IMAGE052
。
Specifically, to calculate the source of uncertainty in soil parameters, a reasonable probability distribution is sought for soil parameters by coefficient of variance (COV). COV is defined as the standard deviation
Figure 952476DEST_PATH_IMAGE053
Ratio of the standard deviation to the standard deviation
Figure 920432DEST_PATH_IMAGE054
The ratio is set to be the same as the input value.
Based on the content of the method embodiment, as an optional embodiment, the newly-built tunnel design optimization method under the construction condition of the proximity tunnel provided in the embodiment of the invention takes the LightGBM regression function as a plurality of objective functions optimized by the NSGA-ii, and simultaneously considers the uncertainty of the geotechnical conditions and the error of the meta-model; establishing probability constraint to carry out multi-objective optimization by a Monte Carlo simulation method, wherein the method comprises the following steps:
Figure 282143DEST_PATH_IMAGE055
wherein,
Figure 259326DEST_PATH_IMAGE056
is the firstjA function of a constraint that is a function of the constraint,
Figure 390093DEST_PATH_IMAGE057
is a defined probability limit for the probability of,
Figure 528951DEST_PATH_IMAGE058
in order to integrate the learning model function,
Figure 502592DEST_PATH_IMAGE059
for the m-th objective function, the first objective function,
Figure 283466DEST_PATH_IMAGE025
is the first in the data setiThe number of the input variables is changed according to the number of the input variables,
Figure 268739DEST_PATH_IMAGE060
is the minimum value of the ith input variable,
Figure 578498DEST_PATH_IMAGE061
is as followsiMaximum value of each input variable.
Specifically, under the condition that a nearby structure exists, the maximum allowable ground surface settlement can be 10 mm during excavation, and the lower limit is 0mm on the assumption that the lifting effect is not needed. The allowable pressure of the shield surface depends on actual conditions, and the maximum pressure can be 250 kpa. The coverage depth is set to range from 2m to 50m and the horizontal distance is set to range from 1m to 20m, in the monte carlo simulation method, it is assumed that at least 99% of the samples are within this range. Aiming at the condition that the difference between the output values of the LSP and the GSD is large, the min-max normalization technology is adopted to select the best result from the Pareto frontier.
Based on the content of the method embodiment, as an optional embodiment, the newly-built tunnel design optimization method provided in the embodiment of the invention under the condition of close tunnel construction calculates the Pareto frontier of the MOO problem by using NSAG-II through defining probability constraints considering soil properties and meta-model uncertainty. The method comprises the following steps: is shown as
Figure 914801DEST_PATH_IMAGE062
The parent population of the generations is
Figure 109154DEST_PATH_IMAGE063
. Through selection, crossover and mutation, progeny populations will be generated
Figure 948934DEST_PATH_IMAGE064
. Suppose that
Figure 695173DEST_PATH_IMAGE063
And
Figure 518772DEST_PATH_IMAGE065
the size of the NSGA-II is N, and the parent population and the offspring population are combined for sorting to ensure the elite property of the NSGA-II. All members will be sorted first in the non-dominant order and will be picked from the lowest sort to the next generation until all N positions are filled. Due to population size limitations, the last non-dominating set cannot generally be taken. To solve this problem, the members belonging to the previous order are further ranked according to the congestion distance, and better can be passed on to the next generation. This process is repeated until the last generation l, the Pareto front is chosen as
Figure 375870DEST_PATH_IMAGE066
。
Based on the content of the method embodiment, as an optional embodiment, in the newly-built tunnel design optimization method under the condition of close tunnel construction provided in the embodiment of the present invention, only one set of optimal results is provided according to the obtained Pareto front edge, and the Pareto optimal solutions are ranked. The method comprises the following steps: and selecting the optimal point with the shortest distance from the ideal point. For the MOO problem, there is usually an ideal point consisting of the extrema of all targets. Will be firstiThe ideal value of each target is represented as
Figure 70156DEST_PATH_IMAGE067
Front of paretojThe distance of the solution to the ideal point is defined as:
Figure 377510DEST_PATH_IMAGE068
wherein,min order to be the target total number,
Figure 422826DEST_PATH_IMAGE069
is shown asjSecond in Pareto solution
Figure 818035DEST_PATH_IMAGE036
The value of the individual object is,
Figure 897987DEST_PATH_IMAGE070
is as followsiThe desired value of the individual target. To be provided withdThe smallest Pareto solution is the final best result.
Specifically, after successfully finding the Pareto front, the LSP and GSD are converted to values between 0 and 1. In order to simulate the tunnel excavation process, the excavation lengths are set to be 10m, 35m and 60 m. After considering uncertainties from geotechnical properties and metamodels, two possible outcomes are expected since the constraint now is probability, rather than a strict "0" or "1" classification. Some pareto optimal solutions may be discarded and some new points may enter the pareto frontier. Fig. 6 and 7 are the Pareto front and the corresponding optimal position, respectively. To further examine the optimization performance, the percent improvement for 20 test cases was calculated under 3 scenarios. And 3 scenes are set, and the difference of the coverage depth and the contribution of the existing tunnel horizontal distance to the optimization result is calculated. The results assuming percent improvement are shown in figure 10.
Taking the result of excavation length 60m as an example, the closest optimal point to the ideal point falls at the positions of C =12.6m and H =8.7 m. In view of the fact that ground settlement increases with increasing excavation processes, it is always desirable to make decisions in the worst case to maintain safety. In this case, the optimum values for LSP and GSD are 84.21kPa and 5.72mm, respectively.
After the uncertainty of the rock-soil condition and the error of the machine learning model are considered, the Pareto frontier is regenerated, and the optimal solution is selected again. As can be seen from fig. 8, the original Pareto solutions with larger LSPs are discarded. As can be seen from fig. 9, the place where the soil coverage is thicker is no longer an ideal choice. After the excavation length reaches 60m, the optimal C value and the optimal H value are respectively 13.1m and 10.7m, the optimal LSP value is reduced to 70.52kPa, and the optimal GSD value is increased to 7.12 mm. The new ideal point sacrifices the GSD to obtain a lower LSP value than the result under the deterministic assumption. Therefore, considering the uncertainty, those high risk solutions are often abandoned, thereby providing a more conservative solution.
To obtain a lower LSP, the results of the GSD will be sacrificed, taking uncertainty into account. The average improvement rate of GSD is reduced from 2.39% to 0.9%, and the average improvement rate of LSP is increased from 9.67% to 11.03%. Meanwhile, the improvement degree of the GSD is also not obvious (0.9% and 11.03% in the case of digging 60 m) compared to the LSP, which means that the improvement efficiency of the LSP is much higher than the GSD. In the three scenarios, 10m has the lowest optimization (2.17% for LSP consideration of 95% PI), and 35m and 60m have the better optimization (8.76% and 11.03% for LSP consideration of 95% PI). This means that the optimization result is more controlled by the depth of coverage of the tunnel than by the horizontal distance of the existing tunnel. But the optimization process can be further improved, giving better results, considering the horizontal distance to the existing tunnel.
The embodiment of the invention provides a newly-built tunnel design optimization method under the construction condition of a proximity tunnel, which is characterized in that an integrated learning model for predicting ultimate support pressure (LSP) and Ground Surface Deformation (GSD) is established by adopting a lightweight gradient booster (LightGBM) algorithm; optimizing two targets of LSP and GSD by using NSGA-II, and simultaneously considering uncertainty of rock and soil conditions and errors of the meta-model; and establishing probability constraint to perform multi-objective optimization (MOO) through Monte Carlo simulation, generating Pareto leading edges, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion. The method can predict and optimize the excavation damage of the adjacent existing tunnel, and also considers various uncertainties such as geological conditions, meta-models and the like so as to ensure the reliability.
The implementation basis of the various embodiments of the present invention is realized by programmed processing performed by a device having a processor function. Therefore, in engineering practice, the technical solutions and functions thereof of the embodiments of the present invention can be packaged into various modules. Based on such a practical situation, on the basis of the foregoing embodiments, an embodiment of the present invention provides an apparatus for optimizing a newly-built tunnel design under a proximity tunnel construction condition, which is used to execute a method for optimizing a newly-built tunnel design under a proximity tunnel construction condition in the foregoing method embodiments. Referring to fig. 2, the apparatus includes: the first main module is used for constructing a shield construction parameter index system and acquiring the existing data; the second main module is used for constructing and training a LightGBM regression model, and establishing an integrated learning model for predicting ultimate support pressure (LSP) and Ground Surface Deformation (GSD) by adopting a lightweight gradient booster (LightGBM) algorithm according to the existing data set; the third main module is used for optimizing two targets of the LSP and the GSD by using NSGA-II, and simultaneously considering uncertainty of geotechnical conditions and errors of the meta-model; and the fourth main module is used for establishing probability constraint for multi-objective optimization (MOO) through Monte Carlo simulation, generating Pareto leading edges, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a criterion.
According to the newly-built tunnel design optimization device under the construction condition of the close-proximity tunnel, which is provided by the embodiment of the invention, a plurality of modules in the figure 2 are adopted, and an integrated learning model for predicting ultimate support pressure (LSP) and Ground Surface Deformation (GSD) is established by adopting a lightweight gradient booster (LightGBM) algorithm; optimizing two targets of LSP and GSD by using NSGA-II, and simultaneously considering uncertainty of geotechnical conditions and errors of the meta-model; and establishing probability constraint to perform multi-objective optimization (MOO) through Monte Carlo simulation, generating Pareto frontier, and obtaining the optimal position of the new tunnel by taking the point with the shortest distance from the ideal point as a standard. The method can predict and optimize the excavation damage of the adjacent existing tunnel, and also considers various uncertainties such as geological conditions, meta-models and the like so as to ensure the reliability.
It should be noted that, the apparatus in the apparatus embodiment provided by the present invention may be used for implementing methods in other method embodiments provided by the present invention, except that corresponding function modules are provided, and the principle of the apparatus embodiment provided by the present invention is basically the same as that of the apparatus embodiment provided by the present invention, so long as a person skilled in the art obtains corresponding technical means by combining technical features on the basis of the apparatus embodiment described above, and obtains a technical solution formed by these technical means, on the premise of ensuring that the technical solution has practicability, the apparatus in the apparatus embodiment described above may be modified, so as to obtain a corresponding apparatus class embodiment, which is used for implementing methods in other method class embodiments. For example:
based on the content of the foregoing device embodiment, as an optional embodiment, the newly-built tunnel design optimization device provided in the embodiment of the present invention under the condition of approaching tunnel construction further includes: a first sub-module, configured to implement the method for establishing an ensemble learning prediction model for predicting ultimate support pressure (LSP) and surface deformation (GSD) using LightGBM algorithm according to the existing data set, including: and dividing the data sample set, randomly selecting four fifths of samples as a training set, remaining one fifths of samples as a test set, carrying out prediction precision analysis on the LightGBM regression model, and evaluating the prediction accuracy of the LightGBM regression model by adopting Mean Square Error (MSE) and Root Mean Square Error (RMSE).
Based on the content of the foregoing device embodiment, as an optional embodiment, the newly-built tunnel design optimization device provided in the embodiment of the present invention under the condition of approaching tunnel construction further includes: a second sub-module for implementing said quantifying the uncertainty of the meta-model with the prediction horizon (PI) of the regression process. The method comprises the following steps: in a linear regression model where y is described by x, given a sample capacity of n, the corresponding PI can be calculated by,
Figure 986029DEST_PATH_IMAGE071
wherein,
Figure 394007DEST_PATH_IMAGE072
is based on
Figure 592907DEST_PATH_IMAGE073
The predicted value of the regression model is determined,
Figure 527365DEST_PATH_IMAGE074
is composed of
Figure 786308DEST_PATH_IMAGE075
Degree of freedom
Figure 540638DEST_PATH_IMAGE076
The value of t at the significant level is,
Figure 543229DEST_PATH_IMAGE077
in order to be the standard error of the residual error,
Figure 456827DEST_PATH_IMAGE016
and
Figure 152251DEST_PATH_IMAGE078
are respectively asxThe mean and the variance of (a) is,nis the sample volume. It is clear from the above equation that a linear regression model must be established before PI. Taking the data of the original data set as y and the predicted value of the LightGBM regression model as x, a linear regression model with the following format can be trained by using an Ordinary Least Square (OLS) method:
Figure 393876DEST_PATH_IMAGE079
based on the content of the foregoing device embodiment, as an optional embodiment, the newly-built tunnel design optimization device provided in the embodiment of the present invention under the condition of approaching tunnel construction further includes: and the third sub-module is used for realizing the quantification of the uncertainty from the geotechnical conditions by using the Monte Carlo simulation technology. The method comprises the following steps: assuming that the soil parameters under consideration obey a certain probability distribution, it will result
Figure 934579DEST_PATH_IMAGE080
The samples obey this distribution. In predicting the input variable as
Figure 843629DEST_PATH_IMAGE018
The algorithm calculates the output value with each sample. All results are within the defined range, which is expressed asmThus predicting a probability within a defined range of
Figure 178795DEST_PATH_IMAGE081
。
Based on the content of the foregoing device embodiment, as an optional embodiment, the device for optimizing design of a newly-built tunnel under a condition of approaching tunnel construction provided in the embodiment of the present invention further includes: the fourth submodule is used for realizing that the LightGBM regression function is used as a plurality of objective functions of NSGA-II optimization, and simultaneously considering uncertainty of geotechnical conditions and errors of the meta-model; through Monte Carlo simulation, probability constraint is established to carry out multi-objective optimization, and the method comprises the following steps:
Figure 48662DEST_PATH_IMAGE082
wherein,
Figure 393056DEST_PATH_IMAGE056
is the firstjA function of a constraint that is a function of the constraint,
Figure 891034DEST_PATH_IMAGE057
is a defined limit of the probability that,
Figure 928260DEST_PATH_IMAGE083
in order to integrate the functions of the learning model,
Figure 144477DEST_PATH_IMAGE059
for the (m) -th objective function (f),
Figure 26983DEST_PATH_IMAGE025
is the first in the data setiThe number of the input variables is changed,
Figure 769680DEST_PATH_IMAGE060
is the minimum value of the ith input variable,
Figure 977807DEST_PATH_IMAGE061
is as followsiMaximum value of each input variable. Suppose for each target
Figure 415742DEST_PATH_IMAGE084
There is an upper limit
Figure 101938DEST_PATH_IMAGE085
And a lower limit
Figure 840087DEST_PATH_IMAGE086
。
Based on the content of the foregoing device embodiment, as an optional embodiment, the newly-built tunnel design optimization device provided in the embodiment of the present invention under the condition of approaching tunnel construction further includes: and the fifth submodule is used for realizing the Pareto frontier of the MOO problem calculated by using NSAG-II through defining probability constraints considering soil properties and metamodel uncertainty. The method comprises the following steps: is shown as
Figure 94482DEST_PATH_IMAGE062
The parent population of the generations is
Figure 19713DEST_PATH_IMAGE063
. Through selection, crossover and mutation, progeny populations will be generated
Figure 509600DEST_PATH_IMAGE065
. Suppose that
Figure 836676DEST_PATH_IMAGE063
And
Figure 121027DEST_PATH_IMAGE065
the size of (2) is N, and the parent population and the offspring population are combined to carry out sequencing, so that the eligibility of NSGA-II is ensured. All members will be sorted first in the non-dominant order and will be picked from the lowest sort to the next generation until all N positions are filled. Due to population size limitations, the last non-dominating set cannot generally be taken. To solve this problem, the members belonging to the previous level will be further ranked according to the congestion distance, and better can be passed on to the next generation. This process is repeated until the last generation l, the Pareto front is selected as
Figure 799133DEST_PATH_IMAGE087
。
Based on the content of the above device embodiment, as an alternative embodiment, the present inventionThe newly-built tunnel design optimization device provided in the embodiment of the invention under the condition of the construction of the proximity tunnel further comprises: and the sixth sub-module is used for realizing that only one group of optimal results are provided according to the obtained Pareto leading edge, and sequencing the Pareto optimal solution. The method comprises the following steps: and selecting the optimal point with the shortest distance from the ideal point. For the MOO problem, there is usually an ideal point consisting of the extrema of all targets. Will be firstiThe ideal value of each target is represented as
Figure 229063DEST_PATH_IMAGE067
Front of paretojThe distance of the solution to the ideal point is defined as:
Figure 410646DEST_PATH_IMAGE068
wherein,mis a target total number of the cells,
Figure 865898DEST_PATH_IMAGE088
is shown asjSecond in Pareto solution
Figure 31300DEST_PATH_IMAGE036
The value of the individual object is,
Figure 597411DEST_PATH_IMAGE089
is as followsiThe desired value of the individual target. To be provided withdThe smallest Pareto solution is the final best result.
The method of the embodiment of the invention is realized by depending on the electronic equipment, so that the related electronic equipment is necessarily introduced. To this end, an embodiment of the present invention provides an electronic apparatus, as shown in fig. 3, including: the system comprises at least one processor (processor), a communication Interface (communication Interface), at least one memory (memory) and a communication bus, wherein the at least one processor, the communication Interface and the at least one memory are communicated with each other through the communication bus. The at least one processor may invoke logic instructions in the at least one memory to perform all or a portion of the steps of the methods provided by the various method embodiments described previously.
In addition, the logic instructions in the at least one memory may be implemented in software functional units and stored in a computer readable storage medium when sold or used as a stand-alone product. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the method embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on the understanding, the above technical solutions substantially or otherwise contributing to the prior art may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the various embodiments or some parts of the embodiments.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. Based on this recognition, each block in the flowchart or block diagrams may represent a module, a program segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In this patent, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A newly-built tunnel design optimization method under the condition of near tunnel construction is characterized by comprising the following steps:
s1, constructing a tunnel construction parameter index system, and establishing an integrated learning model for predicting ultimate support pressure and earth surface deformation by adopting a LightGBM regression model based on the tunnel construction parameter index system;
s2, optimizing the limit supporting pressure and the earth surface deformation by using uncertainty of the rock-soil conditions and the meta-model error of the integrated learning model as constraints;
s3, probability constraint is established, multi-objective optimization is carried out on the integrated learning model, a Pareto front edge is generated, an optimal solution is selected by taking the point with the shortest ideal point distance as a criterion, and the optimal solution is the optimal position of the new tunnel.
2. The method for optimizing the design of the newly-built tunnel under the condition of the close tunnel construction according to claim 1, wherein the step S1 specifically comprises: the method comprises the steps of collecting an existing data set, dividing the existing data set into a training set and a testing set, adopting the training set to carry out prediction accuracy analysis on the LightGBM regression model, and adopting mean square error and root mean square error to evaluate the accuracy of LightGBM regression model prediction.
3. The method of claim 2, wherein in step S2, the data of the existing data set is used as y, and the predicted value of the LightGBM regression model is used as the predicted value of the y
Figure 62736DEST_PATH_IMAGE001
Training a linear regression model by using a least square method as follows:
Figure 228139DEST_PATH_IMAGE002
in the formula,
Figure 59828DEST_PATH_IMAGE003
an estimated value obtained by substituting x into the regression equation,
Figure 751710DEST_PATH_IMAGE004
in order to be a function of the regression equation,
Figure 377863DEST_PATH_IMAGE005
for existing data setsiThe number of the data is set to be,
Figure 30561DEST_PATH_IMAGE006
to predict the valueiThe number of the data is set to be,
Figure 400363DEST_PATH_IMAGE007
is a predicted value of the regression model,
Figure 556538DEST_PATH_IMAGE008
for the average of the existing data set,iis a sample number.
4. The method of claim 3, wherein in the LightGBM regression model, the uncertainty of the meta-model is quantified by using the prediction interval of the regression process, and the method comprises the following steps: in the linear regression model, a given sample capacity isnAs a sample of an existing data setyPredicted value of LightGBM regression model asxCorresponding prediction intervalPICalculated from the following formula:
Figure 494538DEST_PATH_IMAGE009
wherein,
Figure 103374DEST_PATH_IMAGE010
is based on
Figure 542445DEST_PATH_IMAGE011
The predicted value of the regression model is determined,
Figure 818706DEST_PATH_IMAGE012
is composed of
Figure 786662DEST_PATH_IMAGE013
Degree of freedom
Figure 148373DEST_PATH_IMAGE014
At a significant leveltThe value of the one or more of the one,
Figure 250190DEST_PATH_IMAGE015
in order to be the standard error of the residual error,
Figure 115378DEST_PATH_IMAGE016
and
Figure 519814DEST_PATH_IMAGE017
are respectively asxThe mean and the variance of (a) are,nis the sample volume.
5. The method for optimizing the design of the newly-built tunnel under the construction condition of the close-range tunnel according to claim 1, wherein in the step S2, the Monte Carlo simulation method is adopted to quantify the uncertainty from the geotechnical conditions, and the method comprises the following steps: assuming that the soil parameters under consideration obey a certain probability distribution, it will resultnThe samples are subjected to the distribution, at a prediction input variable of
Figure 368822DEST_PATH_IMAGE018
The Monte Carlo simulation method calculates the output value with each sample, all results are within a defined range, expressed asmThus predicting a probability within a defined range of
Figure 618537DEST_PATH_IMAGE019
。
6. The method of claim 1, wherein in step S3, probability constraints are established and multi-objective optimization is performed on the ensemble learning model, the method includes:
Figure 869390DEST_PATH_IMAGE020
wherein,
Figure 320094DEST_PATH_IMAGE021
is the firstjA function of a constraint that is a function of the constraint,
Figure 390818DEST_PATH_IMAGE022
is a defined limit of the probability that,
Figure 709804DEST_PATH_IMAGE023
in order to integrate the learning model function,
Figure 815164DEST_PATH_IMAGE024
is as followsmAn objective function of the number of the target functions,
Figure 295823DEST_PATH_IMAGE025
is the first in the data setiThe number of the input variables is changed,
Figure 853844DEST_PATH_IMAGE026
is as followsiThe minimum value of the number of the input variables,
Figure 101154DEST_PATH_IMAGE027
is a firstiMaximum value of each input variable.
7. The method of claim 1, wherein the method comprises optimizing the design of a newly constructed tunnel under the conditions of the construction of a close-proximity tunnelCharacterized in that in step S3, by defining probability constraints considering soil properties and meta-model uncertainty in geotechnical conditions, the Pareto frontier of the multi-objective optimization problem is calculated by using NSAG-II, and the method comprises the following steps: is shown astThe parent population of the generations is
Figure 61020DEST_PATH_IMAGE028
Through selection, crossover and mutation, progeny populations will be generated
Figure 243740DEST_PATH_IMAGE029
Suppose that
Figure 430002DEST_PATH_IMAGE030
And
Figure 825211DEST_PATH_IMAGE029
all the members are sorted according to a non-dominant order and selected from the lowest order to the next generation until the members are sorted to the next generationNThe positions are filled, wherein due to the limitation of the population size, the members belonging to the previous order will be further sorted according to the crowding distance, the process is repeated until the last generation, and the Pareto frontier is selected as
Figure 639583DEST_PATH_IMAGE031
；
According to a group of optimal solutions provided by the obtained Pareto frontier, sorting the Pareto optimal solutions, including: selecting the optimal point with the shortest distance to the ideal point, and regarding the multi-objective optimization problem, calculating the optimal pointiThe ideal value of each target is represented as
Figure 993204DEST_PATH_IMAGE032
Then put pareto aheadjThe distance of the solution to the ideal point is defined as:
Figure 260237DEST_PATH_IMAGE033
wherein,mis a target total number of the cells,
Figure 724717DEST_PATH_IMAGE034
is shown asjSecond in Pareto solution
Figure 518229DEST_PATH_IMAGE035
The value of the one or more objects is,
Figure 777172DEST_PATH_IMAGE036
is as followsiThe desired value of the individual target.
8. A newly-built tunnel design optimization system under the construction condition of a proximity tunnel is characterized by comprising the following components: the first main module is used for constructing a shield construction parameter index system and acquiring the existing data; the second main module is used for constructing and training a LightGBM regression model and establishing an integrated learning model for predicting ultimate support pressure and surface deformation by adopting a LightGBM algorithm according to the existing data set; the third main module is used for optimizing two targets of limit supporting pressure and earth surface deformation by using NSGA-II, and simultaneously using uncertainty of rock-soil conditions and meta-model errors of the integrated learning model as constraints; and the fourth main module is used for establishing probability constraint through a Monte Carlo simulation method, performing multi-objective optimization on the integrated learning model, generating a Pareto front edge, and obtaining the optimal position of the new tunnel by taking a point with the shortest distance from an ideal point as a criterion.
9. An electronic device, comprising:
at least one processor, at least one memory, and a communication interface; wherein,
the processor, the memory and the communication interface are in communication with each other;
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1 to 7.
10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 7.
CN202210917562.7A 2022-08-01 2022-08-01 Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel Withdrawn CN114996829A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210917562.7A CN114996829A (en) 2022-08-01 2022-08-01 Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210917562.7A CN114996829A (en) 2022-08-01 2022-08-01 Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel

Publications (1)

Publication Number Publication Date
CN114996829A true CN114996829A (en) 2022-09-02

Family

ID=83021958

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210917562.7A Withdrawn CN114996829A (en) 2022-08-01 2022-08-01 Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel

Country Status (1)

Country Link
CN (1) CN114996829A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116702271A (en) * 2023-05-18 2023-09-05 天津大学 Shield construction parameter multi-objective optimization method for machine learning and improved genetic algorithm

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200257971A1 (en) * 2019-01-11 2020-08-13 Chongqing University Full-linear model for optimal power flow of integrated power and natural-gas system based on deep learning methods
CN112417573A (en) * 2020-12-10 2021-02-26 华中科技大学 Multi-objective optimization method for shield tunneling underneath existing tunnel construction based on GA-LSSVM and NSGA-II

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200257971A1 (en) * 2019-01-11 2020-08-13 Chongqing University Full-linear model for optimal power flow of integrated power and natural-gas system based on deep learning methods
CN112417573A (en) * 2020-12-10 2021-02-26 华中科技大学 Multi-objective optimization method for shield tunneling underneath existing tunnel construction based on GA-LSSVM and NSGA-II

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LIMAO ZHANG: "Multi-objective optimization for limiting tunnel-induced damages considering uncertainties", 《RELIABILITY ENGINEERING AND SYSTEM SAFETY》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116702271A (en) * 2023-05-18 2023-09-05 天津大学 Shield construction parameter multi-objective optimization method for machine learning and improved genetic algorithm

Similar Documents

Publication Publication Date Title
Mohamed et al. History matching and uncertainty quantification: multiobjective particle swarm optimisation approach
Mlakar et al. GP-DEMO: differential evolution for multiobjective optimization based on Gaussian process models
US9069916B2 (en) Model selection from a large ensemble of models
Hutahaean et al. On optimal selection of objective grouping for multiobjective history matching
Guthrie Real options analysis of climate-change adaptation: investment flexibility and extreme weather events
Min et al. Pareto-based multi-objective history matching with respect to individual production performance in a heterogeneous reservoir
CN111612528A (en) Method, device and equipment for determining user classification model and storage medium
CN114969953B (en) Optimized shield underpass tunnel design method and equipment based on Catboost-NSGA-III
EP2646927A1 (en) Systems and methods for reducing reservoir simulator model run time
WO2017011469A1 (en) Ensemble based decision making
WO2016095068A1 (en) Pedestrian detection apparatus and method
Zucco et al. Reverse flood routing in natural channels using genetic algorithm
Li et al. Linear programming-based scenario reduction using transportation distance
Monfared et al. Convex hull ranking algorithm for multi-objective evolutionary algorithms
CN108280289A (en) Bump danger classes prediction technique based on local weighted C4.5 algorithms
Bachmatiuk et al. Analysis of the performance of different implementations of a heuristic method to optimize forest harvest scheduling
Lei et al. A comparative study of six hybrid prediction models for uniaxial compressive strength of rock based on swarm intelligence optimization algorithms
CN114996829A (en) Newly-built tunnel design optimization method and equipment under construction condition of close-proximity tunnel
Kaveh et al. Feasibility of PSO-ANFIS-PSO and GA-ANFIS-GA models in prediction of peak ground acceleration
CN104732067A (en) Industrial process modeling forecasting method oriented at flow object
CN113821983A (en) Engineering design optimization method and device based on proxy model and electronic equipment
CN115186936A (en) Optimal well pattern construction method for oil field based on GNN model
Parvaze et al. Optimization of water distribution systems using genetic algorithms: A review
Ferraro et al. Use of evolutionary algorithms in single and multi-objective optimization techniques for assisted history matching
CN113688424A (en) Personalized differential privacy protection method based on weight social network

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20220902