CN113047859A - Shield tunneling parameter control method based on local Fisher soil layer identification - Google Patents
Shield tunneling parameter control method based on local Fisher soil layer identification Download PDFInfo
- Publication number
- CN113047859A CN113047859A CN202110389442.XA CN202110389442A CN113047859A CN 113047859 A CN113047859 A CN 113047859A CN 202110389442 A CN202110389442 A CN 202110389442A CN 113047859 A CN113047859 A CN 113047859A
- Authority
- CN
- China
- Prior art keywords
- shield tunneling
- tunneling parameter
- parameter data
- soil layer
- shield
- 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.)
- Pending
Links
- 230000005641 tunneling Effects 0.000 title claims abstract description 111
- 239000002689 soil Substances 0.000 title claims abstract description 71
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000010276 construction Methods 0.000 claims abstract description 19
- 238000003062 neural network model Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 238000012850 discrimination method Methods 0.000 abstract description 5
- 238000013528 artificial neural network Methods 0.000 description 8
- 239000006185 dispersion Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Images
Classifications
-
- 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
- E21D9/093—Control of the driving shield, e.g. of the hydraulic advancing cylinders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention provides a shield tunneling parameter control method based on local Fisher soil layer identification, which relates to the technical field of shield tunneling parameter control and comprises the following steps: s1: collecting shield tunneling parameter data of a construction site; s2: performing multi-mode analysis on the tunneling parameters by a local Fisher discrimination method; s3: and predicting and optimizing the analyzed tunneling parameters through a BP neural network model. The method can also identify the soil layer in real time according to the shield tunneling data provided by the construction site, so that reasonable shield tunneling parameters are selected, and the safety of shield construction is improved.
Description
Technical Field
The invention relates to the technical field of shield tunneling parameter control, in particular to a shield tunneling parameter control method based on local Fisher soil layer identification.
Background
The shield tunneling parameters mainly relate to total thrust, cutter torque, cutter rotating speed, cutter penetration, tunneling speed, soil bin pressure and the like. The optimization of the tunneling parameters can improve the tunneling efficiency, reduce the abrasion of a cutter head cutter and effectively avoid the phenomenon of surface subsidence caused by unreasonable shield parameters. The shield tunneling parameters are also influenced by the stratum geology, and the tunneling parameters under different stratum conditions should be correspondingly adjusted. The existing tunneling parameter prediction control method mainly comprises the steps of establishing a mathematical model, and establishing a mathematical prediction model among all tunneling parameters based on model tests or field monitoring data, so that the pertinence is strong. The BP (back propagation) neural network can better establish the nonlinear relation among all tunneling parameters by utilizing the advantage of the nonlinear mapping capability of the BP (back propagation) neural network.
However, the BP (back propagation) neural network is lack of comprehensiveness in predicting tunneling parameters, most of shield construction environments are complex strata, and the relevance between tunneling parameters under the complex strata is different. Therefore, the projection direction of the lfda (local Fisher discrete analysis) needs to be found by calculating the local intra-class and inter-class dispersion matrices of the training data through the local Fisher discriminant method. Projecting the training data and the test data to a projection vector to extract a characteristic vector; and finally, calculating the Euclidean distance between the characteristic vectors, and classifying the shield tunneling parameters by using a KNN (K-nearest neighbor) classifier so as to achieve the identification of the soil layer. The local Fisher discrimination method has real-time performance, can directly discriminate the soil layer property on line through the tunneling parameters, and better provides construction guidance for shield tunneling.
Disclosure of Invention
The invention aims to provide a soil layer identification method for judging shield tunneling parameters based on local Fisher, which can optimize the prediction of the tunneling parameters by combining with BP neural network prediction, so that the range of the tunneling parameters is more accurate.
One of the embodiments of the present specification provides a shield tunneling parameter control method based on local Fisher soil layer identification, including the following steps:
s1: collecting shield tunneling parameter data of a construction site;
s2: performing multi-mode analysis on the tunneling parameters by a local Fisher discrimination method, and identifying a soil layer in real time;
s3: and predicting and optimizing the analyzed tunneling parameters through a BP neural network model.
Further, the step S2 specifically includes:
s21: collecting tunneling parameter data transmitted on a construction site, and carrying out normalization processing on the data through z-score;
s22: taking 80% of processed data as a training set of modeling data, searching the projection direction of an LFDA (Linear frequency division multiple access) by calculating a local intra-class and inter-class dispersion matrix of training tunneling parameter data, projecting the training data and test data onto a projection vector, extracting feature vectors, finally calculating Euclidean distances among the feature vectors, and classifying by using a KNN (K nearest neighbor) classifier;
s23: and identifying the soil layer properties under different tunneling parameters by modal identification and characteristic value extraction, classifying the shield tunneling parameters according to the soil layer properties, and predicting and controlling the shield tunneling parameters under different strata through a BP neural network.
Further, in step S1, the tunneling parameters include a cutter head torque, a cutter head rotation speed, a propulsion stroke upper, a propulsion stroke lower, a propulsion stroke left, a propulsion stroke right, a propulsion speed upper, a propulsion speed lower, a propulsion speed left, a propulsion speed right, a front soil pressure upper, a front soil pressure lower, a front soil pressure left, a front soil pressure right, and a tunneling speed.
Furthermore, the z-score is used for carrying out normalization processing on the data, removing extreme values, reducing the influence of the extreme values on the average value and improving the soil layer identification capability.
Further, the LFDA feature extraction essentially performs modal classification and identification on the shield tunneling parameters, and comprises four links of data acquisition, preprocessing, feature extraction, state identification and classification decision. The separability between different classes is maximized by acquiring generalized eigenvalues of the data classes while the minimum diffusion within the classes is preserved and the local multi-modal properties of each soil layer are effectively preserved by the distance-based weighting matrix.
Further, the KNN classifier takes shield tunneling parameters under different soil layers as training data, extracts an optimal discrimination vector through an LFDA (Linear frequency division multiple access) method, and obtains characteristic vectors of various data in the optimal discrimination vector direction through projection. And projecting various types of data in the test data set to the direction of the optimal identification vector respectively, and finally selecting a separator based on a distance method to realize soil layer identification.
Due to the adoption of the technical scheme, compared with the prior art, the invention has the following advantages and positive effects:
according to the method, the intra-class dispersion among shield tunneling parameters of projection characteristics is reduced and the dispersion among shield tunneling parameters under different soil layers is improved through a local Fisher algorithm and a KNN classifier, so that the modal identification of the shield tunneling parameters is better realized, and the function of soil layer identification is improved. Meanwhile, the LFDA method better solves the multi-modal problem of the shield tunneling parameters, samples of the tunneling parameters under a plurality of soil layers can be almost completely separated, the samples can be well recognized for a composite soil layer, the soil layers can be recognized in time for shield tunneling data provided by a construction site, and therefore reasonable shield tunneling parameters can be selected.
The beneficial effects of the invention at least comprise:
1. according to the method, after the soil layer of shield construction is identified based on the local Fisher discrimination method, the BP neural network is used for predicting and controlling the safety range of shield tunneling parameters under different soil layer properties, and the method is more accurate compared with the method that only a regression model is used for analyzing the tunneling parameters of a construction site;
2. the method not only identifies the soil layer of the existing construction site, but also can identify the soil layer through the real-time shield tunneling parameters, and simultaneously predicts and optimizes the tunneling parameters, so that the safety range of the shield tunneling parameters can be better controlled, and the surface subsidence caused by improper control of the tunneling parameters is reduced.
Drawings
Fig. 1 is a schematic flow chart of a shield tunneling parameter control method based on local Fisher soil layer identification according to the present invention;
fig. 2 is a schematic flow chart of the shield tunneling parameter control method based on local Fisher soil layer identification for displaying the soil layer identification based on the LFDA-KNN algorithm;
fig. 3 is a schematic flow chart of the shield tunneling parameter control method based on local Fisher soil layer identification for showing the soil layer identification process;
fig. 4 is a schematic flow chart of the shield tunneling parameter control method based on local Fisher soil layer identification for displaying the soil layer identification result.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
The method mainly comprises the steps of collecting shield tunneling parameter data collected during shield construction in the subway shield construction process, carrying out normalization preprocessing on the data through z-score, then carrying out multi-mode analysis on shield tunneling parameters through a local Fisher discrimination method on the processed data, and classifying the shield tunneling parameters so as to identify the shield tunneling parameters under different soil layers. And finally, predicting the shield tunneling parameters under different soil layer properties through a BP neural network, thereby summarizing the safe numerical range of the shield tunneling parameters, optimizing the shield tunneling parameters and improving the shield construction safety. The following describes a shield tunneling parameter control process based on soil layer identification according to the present invention with reference to the accompanying drawings.
Referring to fig. 1 and 3, the shield machine monitoring and management method of the present invention mainly includes the following steps:
(1) firstly, shield tunneling parameter data of a shield construction site are acquired, wherein the tunneling parameters comprise cutter torque, cutter rotating speed, advance stroke upper, advance stroke lower, advance stroke left, advance stroke right, advance speed upper, advance speed lower, advance speed left, advance speed right, front soil pressure upper, front soil pressure lower, front soil pressure left, front soil pressure right and tunneling speed. And carrying out normalization pretreatment on the collected data by MATLB and z-score, removing abnormal values, reducing the influence of extreme values on the average value and improving the soil layer identification capability.
(2) With reference to fig. 2, 80% of processed data is used as a training set of modeling data, a divergence matrix between shield tunneling parameters of the training data within a class and between shield tunneling parameters of soil layers with different properties is calculated through local Fisher, a projection direction of an LFDA is found, the training data and test data are projected onto a projection vector, feature vectors are extracted, and finally an euclidean distance between the feature vectors is calculated, and classification is performed through a KNN classifier, so that the multi-modal problem of the shield parameters is solved better, samples of the tunneling parameters of a plurality of soil layers can be almost completely separated, recognition for a composite soil layer can be well performed, and the effect of soil layer recognition is shown in fig. 4.
(3) And predicting shield tunneling parameters under soil layers with different properties by a BP neural network algorithm, and taking the cutter torque, the cutter rotating speed, the propelling speed, the front soil pressure and the tunneling speed as research objects. These tunneling parameters are divided into input variables and output variables. The input variables comprise cutter head torque, cutter head rotating speed, advance stroke upper, advance stroke lower, advance stroke left, advance stroke right, advance speed upper, advance speed lower, advance speed left, advance speed right, front soil pressure upper, front soil pressure lower, front soil pressure left and front soil pressure right. Wherein, the penetration is an external factor reflecting the formation condition; the propulsion speed is an output variable as a response output of the system. Through sample training and learning, a network structure and learning parameters are determined, a nonlinear relation between input and output variables is established, and data input and output are realized. And training the network by using a given training sample set, namely repeatedly adjusting the connection weight coefficient of the network and the threshold value of the neuron so that the network realizes a given input-output mapping relation. And finally, forecasting the tunneling parameters, and optimizing the shield tunneling parameter values under different soil layers, thereby improving the shield construction efficiency.
The invention can not only identify the soil layer property through the tunneling parameter, but also predict the parameters such as the shield tunneling speed and the like through the collected tunneling parameter, improve the shield construction efficiency through optimizing and controlling the tunneling parameter, and simultaneously reduce the shield construction fault caused by improper control of the shield tunneling parameter.
The foregoing is merely a preferred embodiment of the invention, it is to be understood that the invention is not limited to the forms disclosed herein, but is not intended to be exhaustive or to limit the invention to other embodiments, and to various other combinations, modifications, and environments and may be modified within the scope of the inventive concept as expressed herein, by the teachings or the skill or knowledge of the relevant art. And that modifications and variations may be effected by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A shield tunneling parameter control method based on local Fisher soil layer identification is characterized by comprising the following steps:
collecting shield tunneling parameter data of a construction site;
performing multi-mode analysis on the shield tunneling parameter data, and identifying a soil layer in real time;
and predicting shield tunneling parameters under different stratums and optimizing the shield tunneling parameters under different stratums by using a BP neural network model based on the analyzed shield tunneling parameter data.
2. The method of claim 1, wherein the shield tunneling parameter data comprises at least one of cutter head torque, cutter head rotation speed, advance stroke up, advance stroke down, advance stroke left, advance stroke right, advance speed up, advance speed down, advance speed left, advance speed right, front soil pressure up, front soil pressure down, front soil pressure left, front soil pressure right and tunneling speed.
3. The method for controlling the shield tunneling parameter based on local Fisher soil layer identification according to claim 1, wherein the multi-modal analysis is performed on the shield tunneling parameter data, and the real-time identification of the soil layer specifically comprises:
preprocessing the shield tunneling parameter data;
performing characteristic extraction on the processed shield tunneling parameter data;
classifying the shield tunneling parameter data based on the characteristics of the shield tunneling parameter data.
4. The method of claim 3, wherein the pre-processing of the shield tunneling parameter data comprises normalizing the shield tunneling parameter data.
5. The method for controlling the shield tunneling parameter based on the local Fisher soil layer identification according to claim 4, wherein the step of normalizing the shield tunneling parameter data comprises the steps of:
and carrying out normalization processing on the shield tunneling parameter data through z-score, and removing extreme values.
6. The method for controlling the shield tunneling parameter based on the local Fisher soil layer identification according to any one of claims 3-5, wherein the step of performing feature extraction on the processed shield tunneling parameter data comprises the following steps:
at least part of the normalized shield tunneling parameter data is used as training data and testing data of the BP neural network model, the projection direction of the LFDA is found by calculating a local intra-class and inter-class divergence matrix for training the shield tunneling parameter data, the training data and the testing data are projected to the projection direction of the LFDA, and the feature vector is extracted.
7. The method of claim 6, wherein the classifying the shield tunneling parameter data based on the characteristics of the shield tunneling parameter data comprises:
and calculating Euclidean distances among the characteristic vectors, and classifying the shield tunneling parameter data by using a classifier.
8. The method for controlling the shield tunneling parameter based on the local Fisher soil layer identification according to claim 7, wherein the classifier is a KNN classifier.
9. The utility model provides a shield tunnelling parameter control device based on local Fisher soil layer discernment, includes at least one storage medium and at least one treater, its characterized in that:
the at least one storage medium is configured to store computer instructions;
the at least one processor is used for executing the computer instructions to realize the shield tunneling parameter control method based on local Fisher soil layer identification according to any one of claims 1-8.
10. A computer-readable storage medium, wherein the storage medium stores computer instructions, and when the computer instructions in the storage medium are read by a computer, the computer executes the shield tunneling parameter control method based on local Fisher soil layer identification according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110389442.XA CN113047859A (en) | 2021-04-12 | 2021-04-12 | Shield tunneling parameter control method based on local Fisher soil layer identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110389442.XA CN113047859A (en) | 2021-04-12 | 2021-04-12 | Shield tunneling parameter control method based on local Fisher soil layer identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113047859A true CN113047859A (en) | 2021-06-29 |
Family
ID=76519228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110389442.XA Pending CN113047859A (en) | 2021-04-12 | 2021-04-12 | Shield tunneling parameter control method based on local Fisher soil layer identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113047859A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106930770A (en) * | 2017-02-06 | 2017-07-07 | 西安科技大学 | Shield machine shield gap method of estimation based on convolutional neural networks |
| CN108304875A (en) * | 2018-01-31 | 2018-07-20 | 中国科学院武汉岩土力学研究所 | A kind of blast fragmentation size prediction technique based on statistic discriminance classification |
| CN108710940A (en) * | 2017-12-31 | 2018-10-26 | 中交第公路工程局有限公司 | Method based on shield machine running orbit parameter in Neural Network Optimization dust stratum |
| CN108952699A (en) * | 2018-07-30 | 2018-12-07 | 中国地质大学（武汉） | A kind of complicated geological drilling process formation lithology intelligent identification Method |
| CN111562285A (en) * | 2020-06-03 | 2020-08-21 | 安徽大学 | Mine water inrush source identification method and system based on big data and deep learning |
| WO2020224233A1 (en) * | 2019-05-05 | 2020-11-12 | 济南轨道交通集团有限公司 | Construction method for shield tunnels passing underneath viaduct in multi-interval, small-clear-distance and overlapping manner |
| CN112632852A (en) * | 2020-12-14 | 2021-04-09 | 西南交通大学 | Karst area subway tunnel shield tunneling speed prediction method and device |
-
2021
- 2021-04-12 CN CN202110389442.XA patent/CN113047859A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106930770A (en) * | 2017-02-06 | 2017-07-07 | 西安科技大学 | Shield machine shield gap method of estimation based on convolutional neural networks |
| CN108710940A (en) * | 2017-12-31 | 2018-10-26 | 中交第公路工程局有限公司 | Method based on shield machine running orbit parameter in Neural Network Optimization dust stratum |
| CN108304875A (en) * | 2018-01-31 | 2018-07-20 | 中国科学院武汉岩土力学研究所 | A kind of blast fragmentation size prediction technique based on statistic discriminance classification |
| CN108952699A (en) * | 2018-07-30 | 2018-12-07 | 中国地质大学（武汉） | A kind of complicated geological drilling process formation lithology intelligent identification Method |
| WO2020224233A1 (en) * | 2019-05-05 | 2020-11-12 | 济南轨道交通集团有限公司 | Construction method for shield tunnels passing underneath viaduct in multi-interval, small-clear-distance and overlapping manner |
| CN111562285A (en) * | 2020-06-03 | 2020-08-21 | 安徽大学 | Mine water inrush source identification method and system based on big data and deep learning |
| CN112632852A (en) * | 2020-12-14 | 2021-04-09 | 西南交通大学 | Karst area subway tunnel shield tunneling speed prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112016602B (en) | Method, equipment and storage medium for analyzing correlation between power grid fault cause and state quantity | |
| CN111505424A (en) | Large experimental device power equipment fault diagnosis method based on deep convolutional neural network | |
| CN111352971A (en) | Bank system monitoring data anomaly detection method and system | |
| CN110737976B (en) | Mechanical equipment health assessment method based on multidimensional information fusion | |
| CN106482967A (en) | A kind of Cost Sensitive Support Vector Machines locomotive wheel detecting system and method | |
| CN112682273B (en) | Wind turbine generator fault detection method based on cost-sensitive lightweight gradient elevator | |
| CN113806889A (en) | Processing method, device and equipment of TBM cutter head torque real-time prediction model | |
| CN114553591A (en) | Training method of random forest model, abnormal flow detection method and device | |
| CN113431635A (en) | Semi-supervised shield tunnel face geological type estimation method and system | |
| Xu et al. | Detection method of insulator based on single shot multibox detector | |
| CN112207631A (en) | Method for generating tool detection model, method, system, device and medium for detecting tool detection model | |
| CN108763926B (en) | Industrial control system intrusion detection method with safety immunity capability | |
| CN114974306A (en) | Transformer abnormal voiceprint detection and identification method and device based on deep learning | |
| CN114328663A (en) | High-dimensional theater data dimension reduction visualization processing method based on data mining | |
| CN113047859A (en) | Shield tunneling parameter control method based on local Fisher soil layer identification | |
| CN112395168A (en) | Stacking-based edge side service behavior identification method | |
| CN111563886A (en) | Tunnel steel rail surface disease detection method and device based on unsupervised feature learning | |
| CN110622692A (en) | Intelligent identification method and system for running state of sugarcane combine harvester | |
| CN115438402A (en) | Real-time TBM tunneling rock mass condition prediction method and device based on rock-machine mutual feedback perception relation | |
| CN111652253A (en) | Well leakage accident detection early warning method based on big data | |
| CN113033683B (en) | Industrial system working condition monitoring method and system based on static and dynamic joint analysis | |
| CN115883182A (en) | Method and system for improving network security situation element identification efficiency | |
| Shi et al. | The turnout abnormality diagnosis based on semi-supervised learning method | |
| CN114463667A (en) | Small sample learning method based on video identification | |
| CN114066075A (en) | Customer loss prediction method based on deep learning |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210629 |
|
| WD01 | Invention patent application deemed withdrawn after publication |