AU2021101934A4 - Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive - Google Patents
Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive Download PDFInfo
- Publication number
- AU2021101934A4 AU2021101934A4 AU2021101934A AU2021101934A AU2021101934A4 AU 2021101934 A4 AU2021101934 A4 AU 2021101934A4 AU 2021101934 A AU2021101934 A AU 2021101934A AU 2021101934 A AU2021101934 A AU 2021101934A AU 2021101934 A4 AU2021101934 A4 AU 2021101934A4
- Authority
- AU
- Australia
- Prior art keywords
- invert
- displacement
- existing tunnel
- nsga
- construction
- 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.)
- Ceased
Links
- 238000010276 construction Methods 0.000 title claims abstract description 87
- 238000005457 optimization Methods 0.000 title claims abstract description 50
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000006073 displacement reaction Methods 0.000 claims abstract description 114
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000012544 monitoring process Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 25
- 230000006870 function Effects 0.000 claims description 21
- 239000002689 soil Substances 0.000 claims description 18
- 238000012360 testing method Methods 0.000 claims description 13
- 230000005641 tunneling Effects 0.000 claims description 13
- 239000006260 foam Substances 0.000 claims description 6
- 238000007620 mathematical function Methods 0.000 claims description 2
- 238000007637 random forest analysis Methods 0.000 abstract description 9
- 238000011156 evaluation Methods 0.000 abstract description 3
- 238000007418 data mining Methods 0.000 abstract 1
- 230000000875 corresponding effect Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000009412 basement excavation Methods 0.000 description 4
- 238000002790 cross-validation Methods 0.000 description 4
- 230000002068 genetic effect Effects 0.000 description 4
- 230000035772 mutation Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000003066 decision tree Methods 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 3
- 230000001276 controlling effect Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000001771 impaired effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/048—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators using a predictor
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
- G05B19/042—Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
- G05B19/0423—Input/output
- G05B19/0425—Safety, monitoring
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
- G05B19/408—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by data handling or data format, e.g. reading, buffering or conversion of data
- G05B19/4083—Adapting programme, configuration
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH DRILLING; MINING
- E21D—SHAFTS; TUNNELS; GALLERIES; LARGE UNDERGROUND CHAMBERS
- E21D9/00—Tunnels or galleries, with or without linings; Methods or apparatus for making thereof; Layout of tunnels or galleries
- E21D9/06—Making by using a driving shield, i.e. advanced by pushing means bearing against the already placed lining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
Abstract
The present invention discloses a method for multi-objective optimization of
construction parameters for shield construction adjacent to an existing tunnel based
on RF and NSGA-II. The method comprises the following steps: constructing a
construction parameter index system for shield construction adjacent to an existing
tunnel and acquiring real-time monitoring data; constructing and training a random
forest regression model, and predicting invert settlement displacement and invert
horizontal displacement of the adjacent existing tunnel caused by shield construction
based on the real-time monitoring data; constructing an RF-NSGA-II multi-objective
optimization model to minimize the invert settlement displacement and the invert
horizontal displacement; and determining optimal shield construction parameter
combinations that can satisfy the minimum invert settlement displacement and invert
horizontal displacement of the existing tunnel based on a resulting Pareto front
solution set. The method effectively improves the prediction accuracy of the invert
settlement displacement and the invert horizontal displacement of the adjacent
existing tunnel caused by shield construction, and provides a new idea for invert
displacement of the existing tunnel and selection of construction parameters from the
perspective of data mining.
1/5
FIGURES OF THE SPECIFICATION
onstructing a shield constructi n
parameter index system
ulti-objective optimization o
-hieldconstructionparameters
RF-based prediction NSGA-I based
-- ulti-objective optimizat ri"
preprocessing shield fun
c struction arameter ata
Horizontal Settlement
Parameter op n isplacemer t o ectivefunc on
no o)
Evaluation of prediction Mul -objective optimization shield
co traction parameters bated on
utputtingRFpredicti areo opima
r-- --------i ----n- n---i-n- ---n l
Fig. 1
02
07 7
0.5
02 2 0 20 40 60 81 100 120 140 160 I8 200 0 20 40 60 60 100 120 140 160 10 200
I, uu V1 1.1iim a1 <
(a) Invert horizontal displacement (b) Invert settlement displacement
Fig. 2
Description
1/5 FIGURES OF THE SPECIFICATION
onstructing a shield constructi n parameter index system
ulti-objective optimization o -hieldconstructionparameters
RF-based prediction NSGA-I based -- ulti-objective optimizat ri"
preprocessing shield fun c struction arameter ata Horizontal Settlement Parameter op n isplacemer t o ectivefunc on no o)
Evaluation of prediction Mul -objective optimization shield co traction parameters bated on
utputtingRFpredicti areoopima ---- r------ i---- n- n---i-n-- --n l
Fig. 1
07 7
02 0.5
02 2 0 0 20 40 60 81 100 120 140 160 I8 200 20 40 60 60 100 120 140 160 10 200 I, uu V1 1.1iim a1 < (a) Invert horizontal displacement (b) Invert settlement displacement
Fig. 2
Intelligent prediction of shield existing tunnels and multi-objective optimization
control method of construction parameters based on data drive
The present invention relates to the technical field of optimization of shield
construction parameters, in particular to a method for multi-objective
optimization of construction parameters for shield construction adjacent to an
existing tunnel based on RF and NSGA-II.
In order to meet the increasing demand for public transport in cities, it is
necessary to develop and utilize underground space reasonably. The subway
effectively relieves the pressure on the ground transportation system with its
advantages of controllable time, safety and convenience. Shield construction
has become the preferred method in urban subway construction because of its
advantages of little disturbance to strata, high degree of automation, fast
tunneling speed, no noise during construction and no impact on ground traffic.
With the development of underground space, there are more and more
subway tunnels undercrossing adjacent existing tunnels. Mechanical effects
on the existing tunnels caused by construction behaviors during shield
tunneling will lead to longitudinal and transverse deformation of the existing
tunnels, and normal service functions of the existing tunnels will be impaired
when the deformation exceeds the requirements of relevant regulations. The
core objective of shield construction is to effectively control the deformation of
existing tunnels and ensure normal operation of existing tunnels and the safety
of undercrossing construction. Therefore, it is very necessary to reasonably predict and control the deformation of existing tunnels during the shield construction of subway tunnels.
In the past, the deformation under existing subway tunnels was generally
predicted by theoretical formulae, numerical analysis and model tests.
However, among the above methods, the theoretical formulae have great
limitations in dealing with complex problems, and the solution accuracy is low.
The numerical analysis is based on a large number of assumptions, with heavy
workload and large fluctuation in accuracy, and thus cannot achieve real-time
control. The model tests have disadvantages such as high cost and long time,
and cannot accurately describe the influence of engineering geological
conditions and shield construction parameters.
Therefore, it is urgent to develop an objective and accurate method for
predicting invert displacement of existing tunnels and optimizing shield
construction parameters.
The purpose of the present invention is to provide a method for
multi-objective optimization of construction parameters for shield construction
adjacent to an existing tunnel based on RF and NSGA-II to solve existing
problems in the prior art. The method can objectively and accurately predict
invert displacement of the existing tunnel adjacent to shield construction and
optimize shield construction parameters.
In order to achieve the purpose, the technical solution used in the present
invention is as follows:
A method for multi-objective optimization of construction parameters for
shield construction adjacent to an existing tunnel based on RF and NSGA-II, comprising the following steps:
S1. constructing a construction parameter index system for shield
construction adjacent to an existing tunnel and acquiring real-time monitoring
data;
S2. constructing and training an RF regression model, and predicting
invert settlement displacement and invert horizontal displacement of the
adjacent existing tunnel caused by shield construction based on the real-time
monitoring data;
S3. constructing an RF-NSGA-II multi-objective optimization model to
minimize the invert settlement displacement and the invert horizontal
displacement with the RF regression function as a fitness function for
optimizing the invert displacement of the existing tunnel by NSGA-II; and
S4. determining optimal shield construction parameter combinations that
can satisfy the minimum invert settlement displacement and invert horizontal
displacement of the existing tunnel based on a resulting Pareto front solution
set.
Preferably, the construction parameters for shield construction adjacent to
the existing tunnel in the step S1 comprise: soil chamber pressure, foam
amount, grouting amount, tunneling speed, cutterhead torque and jacking
force.
Preferably, the step of constructing and training an RF regression model
in the step S2 comprises:
S2.1. acquiring and preprocessing data: normalizing invert displacement
data of the existing tunnel;
S2.2. optimizing RF parameters: training samples by an RF regression algorithm, and adjusting important parameters in the model;
S2.3. predicting invert displacement: dividing a data sample set, selecting
four fifths of the samples randomly as a training set, taking the remaining one
fifth of the samples as a test set, and optimizing parameters based on the RF
model; and
S2.4. analyzing prediction accuracy: evaluating the prediction accuracy of
the model based on root-mean-square error (RMSE) and goodness of fit R 2
. Preferably, the step of optimizing the invert settlement displacement and
the invert horizontal displacement of the adjacent existing tunnel caused by
shield construction by the RF-NSGA-II model in the step S3 comprises:
S3.1. determining objective functions and constraints: replacing traditional
mathematical functions by an RF regression prediction algorithm for the invert
displacement of the existing tunnel as the objective functions in NSGA-II, and
setting limits for decision variables based on actual conditions of the project
and relevant regulations to form constraints on the variables;
S3.2. multi-objective optimization of invert settlement displacement and
invert horizontal displacement: achieving multi-objective optimization of shield
construction parameters by a NSGA-II algorithm, and determining a Pareto
optimal solution set of shield construction parameters for a new tunnel
undercrossing the existing tunnel to ensure the safety of invert displacement of
the existing tunnel.
Preferably, optimal shield construction parameter combinations that can
satisfy the minimum invert settlement displacement and invert horizontal
displacement of the existing tunnel are determined based on the Pareto
optimal solution set, and optimal solutions are obtained by an ideal point method.
Preferably, the step of obtaining optimal solutions by an ideal point
method comprises:
S4.1. determining an ideal point for optimizing problems based on optimal
values for objectives in the Pareto optimal solution set;
S4.2. calculating the distance d, between each Pareto optimal solution
and the ideal point; and
S4.3. selecting the Pareto solution corresponding to the minimum d, as
the optimal compromise solution based on the principle of minimum distance.
The present invention has the following advantageous effects:
(1) The present invention provides a method for multi-objective
optimization by an RF-NSGA-II intelligent algorithm, which achieves the
objective of reducing invert settlement displacement and invert horizontal
displacement of an existing tunnel in the project by optimizing shield
construction parameters such as soil chamber pressure, foam amount,
synchronous grouting amount, tunneling speed, cutterhead torque and jacking
force.
(2) Data samples are used in the present invention to establish the
nonlinear mapping relationship between the invert settlement displacement
and the invert horizontal displacement of the adjacent existing tunnel and the
shield construction parameters by applying an RF algorithm, so as to improve
the prediction accuracy. The multi-objective optimization of the invert
settlement displacement and the invert horizontal displacement of the adjacent
existing tunnel is carried out through the combination of RF and NSGA-II, and
relevant factors to be considered in shield construction are modeled more comprehensively, which helps to get a more reasonable design scheme for construction parameters.
In order to explain more clearly the embodiments in the present invention
or the technical solution in the prior art, the following will briefly introduce the
figures needed in the description of the embodiments. Obviously, figures in the
following description are only some embodiments of the present invention, and
for a person skilled in the art, other figures may also be obtained based on
these figures without paying any creative effort.
Fig. 1 is a flowchart of a multi-objective optimization model for shield
construction parameters based on RF-NSGA-II algorithm.
Fig. 2 is a schematic diagram of optimization results of random forest
parameters in an embodiment of the present invention.
Fig. 3 is a schematic diagram of prediction results of settlement
displacement of tunnel invert in an embodiment of the present invention.
Fig. 4 is a schematic diagram of prediction results of horizontal
displacement of tunnel invert in an embodiment of the present invention.
Fig. 5 is a flowchart of the NSGA-II algorithm of the present invention.
Fig. 6 shows a Pareto ideal point of the present invention.
Fig. 7 is a schematic diagram of a Pareto front solution set in an
embodiment of the present invention.
The technical solution in the embodiments of the present invention will be
described clearly and completely with reference to the figures in the
embodiments of the present invention; obviously, the described embodiments are only part of the embodiments of the present invention and not all of them.
Based on the embodiments of the present invention, all other embodiments
obtained by those of ordinary skill in the art without creative work should fall
within the protection scope of the present invention.
The present invention will be further described in detail with reference to
accompanying drawings and preferred embodiments for clear understanding
of the above purpose, features and advantages of the present invention.
As shown in Fig. 1, the present invention provides a method for
multi-objective optimization of construction parameters for shield construction
adjacent to an existing tunnel based on RF and NSGA-II, comprising the
following steps:
S1. constructing a construction parameter index system for shield
construction adjacent to an existing tunnel and acquiring real-time monitoring
data, which specifically comprises:
S1.1. a construction parameter index system for shield construction
adjacent to the existing tunnel, including:
soil chamber pressure: soil chamber pressure is an important guarantee
for excavation face stability during shield tunneling, and the reasonable control
of soil chamber pressure is crucial to reduce soil disturbance in front of the
excavation face;
foam amount: the rotation of cutterhead and the propulsion of shield
tunneling machine will cause soil deformation. Therefore, foam is usually
injected to improve soil properties during construction, so as to reduce
cutterhead torque of the shield tunneling machine, and maintain relative
stability of soil, thereby reducing the disturbance to strata; grouting amount: the purpose of synchronous grouting is mainly to fill tail voids in time to integrate segments with the surrounding soil and reduce the friction between the segments and the soil; tunneling speed: tunneling speed is directly related to the cutterhead speed, the greater the cutterhead speed, the greater the disturbance to the soil in front of the excavation face, and vice versa; cutterhead torque: cutterhead torque is to balance the friction torque between cutterhead panel and soil at excavation face, the torque of cutting soil with cutter, and the torque generated by mixing muck in soil chamber; and jacking force: jacking force is mainly to overcome friction resistance and resistance to cutterhead to propel shield tunneling machine, and has great influence on stratum displacement;
S1.2. acquiring the corresponding monitoring data for construction
parameters for shield construction adjacent to the existing tunnel.
S2. Predicting invert settlement displacement and invert horizontal
displacement of the adjacent existing tunnel caused by shield construction by
an RF regression model, which specifically comprises the following steps:
S2.1. acquiring and preprocessing data: the data were derived from actual
monitoring results of the project, since the six input variables had different
dimensional and attribute ranges and could not be compared directly, it was
necessary to preprocess the data for normalization.
Data preprocessing was mainly carried out to normalize invert
displacement data of the existing tunnel so as to avoid some data in the
samples being too large or too small, which increased the burden on the
algorithm during training and resulted in data flooding or network nonconvergence. Data normalization allowed input data to be in a certain interval, such as [0,1], [-1,1], eliminating the influence of eigenvalue dimensions of different samples on prediction efficiency and accuracy. Since normalization of the data to the interval [-1,1] avoided flooding the features of input vectors better than the interval [0,1], the sample input data were normalized to the interval [-1,1] in the present invention. According to the present invention, input variables and output invert displacement were normalized to the interval [-1,1] by the formula (1) to achieve data normalization and unification of variable dimensions, allowing all features to take effect in the prediction.
Y= (Ymax- .)*min +min (1) Emax EminI
where y was a normalized standard value, ymax and ymin were 1 and -1 by
default respectively, x was the sample value, Xmax and Xmin were the maximum
sample value and the minimum sample value respectively.
S2.2. Optimizing RF parameters;
The setting of RF parameters will directly affect the regression fitting
performance of the model. Therefore, it is necessary to adjust important
parameters of the model first when training samples by random forest
regression algorithm. RandomForestRegressor (RF) parameters to be
adjusted mainly include two parts, parameters of Bagging framework and
parameters of CART decision tree. The most important parameter of the RF
bagging framework is the maximum iterations n_estimators of a weak learner,
the value has a direct impact on the prediction performance of the RF model. If
the value is too small, under-fitting occurs easily, if the value is too large, over-fitting occurs easily. Important parameters of the CART decision tree include maxfeatures and max-depth, which will affect the establishment of the decision tree model, thus affecting the regression fitting performance of the model.
At present, the methods used to optimize model parameters mainly
include enumeration, grid search, particle swarm optimization and genetic
algorithm. Among them, grid search is a global search method for parameter
optimization by traversing given parameter combinations, by which global
optimal solutions can be obtained, and the time taken to find the optimal
solutions is less than enumeration. Therefore, parameters of the RF model are
optimized and adjusted by grid search in the present invention.
In order to determine a set of optimal parameters from all possible
parameter combinations, the performance of the random forest model is
validated by K-fold cross-validation (K-CV) in the present invention to find
parameters corresponding to the model with the highest accuracy and
determine the parameters as the final optimal parameters. K-CV is the most
commonly used form of cross-validation, which can not only avoid
under-learning state, but also avoid over-learning state, and the established
model can also get ideal results when the original sample data size is relatively
large.
The basic idea of K-CV is to divide the original sample data set into K
sample subsets, where K-1 sample subsets are used as a training set and the
remaining sample subset is used as a test set for repeated cross-validation
until every sample subset has a test set.
In the embodiment, a random forest training model was established from the input training sample set by using sklearn modules of python language. In order to avoid the over-fitting state of the random forest model and improve the prediction accuracy and generalization ability of the model, it was necessary to optimize the important parameters maxfeatures, n_estimators and max-depth. Since there were only 6 features selected for training samples, the maxfeatures could be auto by default, while for the optimization of the hyper-parameter n_estimators by grid search and 5-fold cross-validation, the goodness of fit R 2 was taken as a performance evaluation index. The n_estimators was set at 1-200, with an increasing step of 10, and the max-depth was set at 5-8 based on the sample size, with an increasing step of
1. The two parameters were combined to establish a model, with optimization
results shown in Fig. 2. According to code calculation, the combined values of
n_estimators and maxdepth which minimized the prediction model error in
invert settlement displacement were maxdepth=7, n-estimators=75 and
R 2=0.9676, while the combined values of n_estimators and max-depth which
minimized the prediction model error in invert horizontal displacement were
max-depth=6, n_estimators=48 and R 2 =0.9685.
S2.3. Predicting invert displacement;
In the embodiment of the present invention, the training set was used for
learning simulation based on the optimization results of RF parameters to
establish an RF prediction model for settlement displacement and an RF
prediction model for horizontal settlement displacement of tunnel invert
respectively, and the prediction models of the training set were tested by a test
set, with prediction results shown in Fig. 3 and Fig. 4 respectively.
S2.4. Analyzing prediction accuracy;
As can be seen from Fig. 3 and Fig. 4, the settlement displacement and
horizontal displacement of tunnel invert can be predicted by the RF model well.
Fig. 3(a) shows that the RF model has fully learnt the laws between input
variables and output indexes of the training sample set. The results from
prediction of invert settlement displacement by the training set were basically
consistent with the actual observed values. According to the calculation, the
MSE was 0.0198, and the goodness of fit R 2 was 0.9510. Fig. 3(b) shows the
comparison of prediction results by the test set, which is intended to validate
the modeling effect of the training set. It could be seen that the predicted
values of the invert settlement displacement by the test set were highly
consistent with the actual observed values. According to the calculation, the
MSE was 0.0045 and the goodness of fit R 2 was 0.9920. The results from both
the training set and the test set indicated that the training RF prediction model
had high precision and good generalization performance, and the fitted
nonlinear prediction function had high accuracy.
Figs. 4(a) and 4(b) show that the predicted values from the RF prediction
model for invert horizontal displacement on the training set and the test set are
highly consistent with actual values. According to the calculation, the MSE of
the training set was 0.0063, the goodness of fit R 2 was 0.9444, the MSE of the
test set was 0.0025, and the goodness of fit R 2 was 0.9673, which achieved
high accuracy. The results shown in Fig. 3 and Fig. 4 demonstrate that RF is
applicable to the prediction of tunnel invert displacement. To visualize the
decision-making process of random forest in predicting invert settlement
displacement and invert horizontal displacement, an svg vector diagram is
drawn by python for visual representation.
S3. Optimizing invert settlement displacement and invert horizontal
displacement of the adjacent existing tunnel caused by shield construction by
an RF-NSGA-II model, which specifically comprises the following steps:
S3.1. determining objective functions and constraints:
The invert settlement displacement objective function min
g1:min(RF(xi,x2-,x)) based on random forest and the invert horizontal
displacement objective function min g2: min(RF(x,x2 ,---,x6 )) based on
random forest could be obtained by introducing trained RF prediction
regression functions of invert settlement displacement and invert horizontal
displacement as objective functions for NSGA-II global optimization; where x1
, x2, X 3 , X4 , Xs and x6 represented soil chamber pressure, foam amount,
synchronous grouting amount, tunneling speed, cutterhead torque and jacking
force respectively.
Constraints are set for all influencing factor parameters based on actual
conditions of the project. When NSGA-II algorithm is used to search for the
optimal solutions for shield construction parameters, the decision region of
initial population should be set first to ensure that the initial population has
practical significance. In order to avoid potential safety hazards due to large
span of parameter adjustment caused by large difference between optimized
parameters and actual engineering parameters of the shield tunneling machine,
the scope of initial decision variables are intended to be set based on
fluctuations in shield parameters from selected samples as the main reference
in the present invention.
235: x: 390 10: x2 : 30 5 5:x,3.:151 s.t::! 10 : x 4 < 45 1200: x 5: 4266 12 0 0 0 : x,: 19400
S3.2. Multi-objective optimization of invert settlement displacement and
invert horizontal displacement;
In the embodiment of the present invention, the NSGA-II algorithm was
applied to achieve multi-objective optimization of shield construction
parameters, so as to determine the Pareto optimal solution set of shield
construction parameters of a new tunnel undercrossing an existing tunnel and
ensure the safety of invert displacement of the existing tunnel. The
implementation process of the NSGA-II algorithm is shown in Fig. 5, and its
basic idea is as follows:
(1) An initial population P with a size of N is generated randomly, then fast
non-dominated sorting is performed on the initial population, and selection,
crossover and mutation operations of genetic algorithms are realized to
generate an offspring population Q with a size of N;
(2) merging the parent population P and the offspring population Q into a
new population R with a population size of 2N, fast non-dominated sorting and
congestion calculation are performed on the population R, and the first N
individuals in the sorting are selected to form a new parent population;
(3) A new offspring population is generated through the selection,
crossover and mutation operations of genetic algorithms; and
(4) The above procedures are repeated until conditions for end of program
are met to output the Pareto optimal solution set.
S4. obtaining optimal solutions by an ideal point method, which
specifically comprises:
The Pareto optimal solutions obtained by NSGA-II algorithm are a set of
solutions conforming to Pareto optimal state decision variables, but not a
unique solution. In order to select an optimal solution from the Pareto front
solution set as the final decision for practical engineering application, the
present invention adopts the ideal point method to obtain an optimal
compromise solution, that is, the solution closest to the ideal point is selected
as the optimal compromise solution, and the implementation process is as
follows:
(1) determining an ideal point for optimizing problems based on optimal
values for objectives in the Pareto front solution set, and representing the
Pareto optimal solution set and the ideal point obtained by NSGA-II algorithm
by points in Fig. 6 to make the process more intuitive, with the point E in the
figure indicating the ideal point.
(2) Defining the ideal point as f* = (f* ,f2f and the Pareto optimal
solution set as f" = (f",f2"), and calculating the distance between each
Pareto optimal solution and the ideal point as follows:
I+ 2 (2) d=
where, f;*,f* are ideal values corresponding to an objective function 1
and an objective function 2 of the ideal point respectively, and f",f2" are
values corresponding to an objective function 1 and an objective function 2 of
the nth Pareto optimal solution f".
(3) Selecting the Pareto solution corresponding to the minimum d, as the
optimal compromise solution based on the principle of minimum distance.
In the embodiment of the present invention, the NSGA-II algorithm was
applied to search for the optimal solutions for global construction parameter
combinations with the two optimization objectives of controlling the invert
horizontal displacement and settlement displacement, and the number of
objectives, population size, crossover and mutation operators and criteria for
end of optimization of the genetic algorithms were determined before NSGA-II
multi-objective optimization. Considering that proper population size and
iterations can promote the convergence of multi-objective optimization, in the
NSGA-IIalgorithm, the crossover operator was 0.7, the mutation operator was
0.09, the population size was 50, the generations and stallGenLimit were 80 in
the present invention. After the parameters were set, the NSGA-II algorithm
was run to obtain a Pareto front solution set, as shown in Fig. 7, including 50
pairs of optimal solutions.
There was a conflict in the optimization of invert horizontal displacement
and settlement displacement. As can be seen from Fig. 7, the absolute values
of horizontal displacement and settlement displacement were negatively
correlated, and the settlement displacement gradually increased as the invert
horizontal displacement decreased, indicating that the two objective functions
could not be achieved at the same time. However, in the process, the two
displacements would reach a relatively balanced state. According to the
Pareto optimal solution set, the maximum horizontal displacement of tunnel
invert was 1.80mm and the maximum settlement displacement of tunnel invert
was 5.98mm, the values were significantly improved compared with the mean values of horizontal displacement and settlement displacement of 2.09mm and
6.46 mm in the original data samples, indicating that the solutions obtained by
the NSGA-II algorithm could achieve the dual objectives of controlling and
reducing the horizontal displacement and settlement displacement of tunnel
invert at the same time. However, a decision-making process was required to
obtain satisfactory results of the two optimization objectives.
The coordinates of the ideal point formed when the invert horizontal
displacement and settlement displacement reached an optimal state were E
(1.55,4.18). After obtaining the coordinates of the ideal point, 40 points in the
Pareto front solution set were substituted into a distance function formula (2) to
calculate the distance between each optimal solution and the ideal point
respectively, and points with the minimum distance were taken as the optimal
solutions for multi-objective optimization of shield construction parameters. To
validate the effect of multi-objective optimization, the shield construction
parameters as well as the mean values of the invert horizontal displacement
and the invert settlement displacement detected by the samples were
compared with the optimized values, as shown in Table 1.
Table 1
Soil FoamSynchronousTu chamber am grouting nnelling Cutterhead Jacking Invert horizontal Invert settlers pressure arnun amount /(rm/min) torque/(kN-m)force/(kN)displacement/(mm)displacement/i /(kPa) /(m3) Maximum 384 30 14 45 4181 19367 3 9.2 value II Minimum 238 10 5.3 10 12070 12010 1.2 3.3 value Mean 305.28 19.86 9.76 26.60 2694.60 15613.87 2.09 6.46 value Optimized 358.02 18.53 10.30 10.87 1624.80 13047.31 1.73 4.42 value befoundthatthehorizontaldisplacementandsettlement
It can be found that the horizontal displacement and settlement displacement of tunnel invert were significantly decreased after optimization based on the NSGA-II algorithm. The mean horizontal displacement of tunnel invert collected from monitoring samples during the original construction was
2.09mm, after optimization of the shield construction parameters, the
horizontal displacement was 1.73mm, which was 17.2% less than the mean
invert horizontal displacement of the existing tunnel of test samples for original
construction. Meanwhile, the invert settlement displacement was also greatly
improved after optimization. The mean invert settlement displacement of
monitoring samples for original construction was 6.46mm, and the optimized
invert settlement displacement was 4.42mm, which was 31.6% less than the
mean invert settlement displacement of monitoring samples for original
construction, validating that the invert horizontal displacement and settlement
displacement of the existing tunnel caused by shield construction were
obviously improved after multi-objective optimization.
In the description of the present invention, it should be noted that the
orientation or position relationship indicated by the terms "longitudinal",
"transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal",
"top", "bottom", "inside", "outside" and the like is based on the orientation or
position relationship shown in the figures, which is intended only to facilitate
the description of the present invention instead of indicating or implying that
the indicated device or element must have a specific orientation, and be
constructed and operated in a specific orientation, and thus should not be
understood as a limitation to the present invention.
The preferred embodiments described herein are only for illustration
purpose, and are not intended to limit the present invention. Various modifications and improvements on the technical solution of the present invention made by those of ordinary skill in the art without departing from the design spirit of the present invention shall fall within the protection scope as claimed in claims of the present invention.
Claims (1)
1. A method for multi-objective optimization of construction parameters for
shield construction adjacent to an existing tunnel based on RF and NSGA-II,
characterized by comprising the following steps:
S1. constructing a construction parameter index system for shield
construction adjacent to an existing tunnel and acquiring real-time monitoring
data;
S2. constructing and training an RF regression model, and predicting
invert settlement displacement and invert horizontal displacement of the
adjacent existing tunnel caused by shield construction based on the real-time
monitoring data;
S3. constructing an RF-NSGA-II multi-objective optimization model to
minimize the invert settlement displacement and the invert horizontal
displacement with the RF regression function as a fitness function for
optimizing the invert displacement of the existing tunnel by NSGA-II; and
S4. determining optimal shield construction parameter combinations that
can satisfy the minimum invert settlement displacement and invert horizontal
displacement of the existing tunnel based on a resulting Pareto front solution
set.
2. The method for multi-objective optimization of construction parameters
for shield construction adjacent to an existing tunnel based on RF and NSGA-II
according to claim 1, characterized in that the construction parameters for
shield construction adjacent to the existing tunnel in the step S1 comprise: soil
chamber pressure, foam amount, grouting amount, tunneling speed,
cutterhead torque and jacking force.
3. The method for multi-objective optimization of construction parameters
for shield construction adjacent to an existing tunnel based on RF and NSGA-II
according to claim 1, characterized in that the step of constructing and training
an RF regression model in the step S2 comprises:
S2.1. acquiring and preprocessing data: normalizing invert displacement
data of the existing tunnel;
S2.2. optimizing RF parameters: training samples by an RF regression
algorithm, and adjusting parameters in the model;
S2.3. predicting invert displacement: dividing a data sample set, selecting
four fifths of the samples randomly as a training set, taking the remaining one
fifth of the samples as a test set, and optimizing parameters based on the RF
model; and
S2.4. analyzing prediction accuracy: evaluating the prediction accuracy of
the model based on root-mean-square error (RMSE) and goodness of fit R 2
. 4. The method for multi-objective optimization of construction parameters
for shield construction adjacent to an existing tunnel based on RF and NSGA-II
according to claim 1, characterized in that the step of constructing an
RF-NSGA-II multi-objective optimization model in the step S3 comprises:
S3.1. determining objective functions and constraints: replacing traditional
mathematical functions by an RF regression prediction algorithm for the invert
displacement of the existing tunnel as the objective functions in NSGA-II, and
setting limits for decision variables based on actual conditions of the project
and relevant regulations to form constraints on the variables;
S3.2. multi-objective optimization of invert settlement displacement and
invert horizontal displacement: achieving multi-objective optimization of shield construction parameters by a NSGA-II algorithm, and determining a Pareto optimal solution set of shield construction parameters for a new tunnel undercrossing the existing tunnel to ensure the safety of invert displacement of the existing tunnel.
5. The method for multi-objective optimization of construction parameters
for shield construction adjacent to an existing tunnel based on RF and NSGA-II
according to claim 4, characterized in that optimal shield construction
parameter combinations that can satisfy the minimum invert settlement
displacement and invert horizontal displacement of the existing tunnel are
determined based on the Pareto optimal solution set, and optimal solutions are
obtained by an ideal point method.
6. The method for multi-objective optimization of construction parameters
for shield construction adjacent to an existing tunnel based on RF and NSGA-II
according to claim 5, characterized in that the step of obtaining optimal
solutions by an ideal point method comprises:
S4.1. determining an ideal point for optimizing problems based on optimal
values for objectives in the Pareto optimal solution set;
S4.2. calculating the distance d, between each Pareto optimal solution
and the ideal point; and
S4.3. selecting the Pareto solution corresponding to the minimum d, as
the optimal compromise solution based on the principle of minimum distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2021101934A AU2021101934A4 (en) | 2021-04-15 | 2021-04-15 | Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| AU2021101934A AU2021101934A4 (en) | 2021-04-15 | 2021-04-15 | Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| AU2021101934A4 true AU2021101934A4 (en) | 2021-06-03 |
Family
ID=76132897
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| AU2021101934A Ceased AU2021101934A4 (en) | 2021-04-15 | 2021-04-15 | Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive |
Country Status (1)
| Country | Link |
|---|---|
| AU (1) | AU2021101934A4 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114418469A (en) * | 2022-03-30 | 2022-04-29 | 华中科技大学 | LGBM-NSGA-III-based shield proximity construction parameter multi-objective optimization method and device |
| CN114818288A (en) * | 2022-04-14 | 2022-07-29 | 中铁建电气化局集团南方工程有限公司 | Method for constructing subway three-dimensional trackless measuring platform |
| CN114996830A (en) * | 2022-08-03 | 2022-09-02 | 华中科技大学 | Visual safety assessment method and equipment for shield tunnel to pass through existing tunnel |
-
2021
- 2021-04-15 AU AU2021101934A patent/AU2021101934A4/en not_active Ceased
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114418469A (en) * | 2022-03-30 | 2022-04-29 | 华中科技大学 | LGBM-NSGA-III-based shield proximity construction parameter multi-objective optimization method and device |
| CN114818288A (en) * | 2022-04-14 | 2022-07-29 | 中铁建电气化局集团南方工程有限公司 | Method for constructing subway three-dimensional trackless measuring platform |
| CN114818288B (en) * | 2022-04-14 | 2023-05-09 | 中铁建电气化局集团南方工程有限公司 | Construction method of subway three-dimensional trackless measurement platform |
| CN114996830A (en) * | 2022-08-03 | 2022-09-02 | 华中科技大学 | Visual safety assessment method and equipment for shield tunnel to pass through existing tunnel |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2021101934A4 (en) | Intelligent prediction of shield existing tunnels and multi objective optimization control method of construction parameters based on data drive | |
| AU2021101920A4 (en) | Multi-objective intelligent optimization control method of shield construction parameters based on data mining | |
| CN105139585B (en) | A kind of soil-slope dangerous situation intelligent early-warning forecasting procedure | |
| CN103984788B (en) | A kind of coal entry anchor rod support automated intelligent design and optimization system | |
| CN105260575A (en) | Roadway surrounding rock deformation predicting method based on neural network | |
| Deng et al. | Risk evaluation model of highway tunnel portal construction based on BP fuzzy neural network | |
| Miranda et al. | Geomechanical characterization of volcanic rocks using empirical systems and data mining techniques | |
| Wang et al. | Risk Assessment of Water Inrush in Karst Tunnels Based on the Efficacy Coefficient Method. | |
| CN114969953B (en) | Optimized shield underpass tunnel design method and equipment based on Catboost-NSGA-III | |
| Chen et al. | Prediction of TBM tunneling parameters through an LSTM neural network | |
| CN112879024A (en) | Dynamic prediction method, system and equipment for shield attitude | |
| CN115423167A (en) | Subway deep foundation pit construction safety early warning and decision-making assisting method and system | |
| Adoko et al. | Estimation of convergence of a high-speed railway tunnel in weak rocks using an adaptive neuro-fuzzy inference system (ANFIS) approach | |
| CN114357750A (en) | Goaf water filling state evaluation method | |
| CN113323676B (en) | Method for determining cutter torque of shield machine by using principal component analysis-long and short memory model | |
| CN113033954B (en) | Intelligent decision-making method for coordinated development of coal and coalbed methane | |
| Bouayad et al. | Assessment of ground surface displacements induced by an earth pressure balance shield tunneling using partial least squares regression | |
| Sun et al. | Inversion of Surrounding Rock Mechanical Parameters in a Soft Rock Tunnel Based on a Hybrid Model EO-LightGBM | |
| Fomychov et al. | Determination of parameters of non-uniform fractured rock massif in computing experiment | |
| CN117113842A (en) | Tunnel surrounding rock deformation prediction method based on PSO-LSTM model | |
| Mahdevari et al. | A hierarchical local-model tree for predicting roof displacement in longwall tailgates | |
| CN110555598A (en) | Fuzzy comprehensive evaluation method for stability of karst foundation | |
| Qu et al. | Mining Engineering Image Recognition Method Based on Simulated Annealing Algorithm | |
| Liu et al. | Curve forming prediction of coal mine roadway based on RBF interpolation | |
| Yue et al. | Influencing Factors and Adaptability of Working Face Length in Deep Thick Coal Seam |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FGI | Letters patent sealed or granted (innovation patent) | ||
| MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |