CN110006348B - Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel - Google Patents
Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel Download PDFInfo
- Publication number
- CN110006348B CN110006348B CN201910209178.XA CN201910209178A CN110006348B CN 110006348 B CN110006348 B CN 110006348B CN 201910209178 A CN201910209178 A CN 201910209178A CN 110006348 B CN110006348 B CN 110006348B
- Authority
- CN
- China
- Prior art keywords
- data
- segment
- joint
- waterproof
- optical fiber
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Lining And Supports For Tunnels (AREA)
Abstract
The invention relates to an early warning type intelligent monitoring method for waterproof performance of a segment joint of a subway shield tunnel, which is used for monitoring the waterproof performance of the segment joint.
Description
Technical Field
The invention belongs to the field of shield tunnel engineering, and particularly relates to an intelligent monitoring method for waterproof performance of a subway shield tunnel joint.
Background
In the subway shield tunnel operation process, the percolating water problem of section of jurisdiction seam position has adverse effect to tunnel structure health, the durability and the suitability of tunnel can be weakened, if administer in time not, on the one hand the ring outside infiltration causes the soil particle loss on every side, cause the change of ring outside stress field, produce certain adverse effect to the ring structure, on the other hand the seam infiltration permeates inside the section of jurisdiction seam through the sealing gasket back, under certain carbon dioxide concentration and relative humidity environment, cause the carbonization of joint concrete easily, along with the carbonization degree of depth constantly increases, the corruption of reinforcing bar and bolt is accelerated, the rigidity that the section of jurisdiction connects has been reduced. Therefore, in order to guarantee the normal use in the subway shield tunnel, waterproof monitoring to the shield tunnel seam is indispensable.
At present, the traditional shield tunnel water leakage monitoring method in China is to arrange 4 persons with professional qualifications to carry out periodic detection on the shield tunnel, wherein: the method comprises the steps of lighting, photographing, detecting tunnels and recording, wherein 1 person is arranged respectively, under the common condition, the time is required to be about 3h to 4h when 1 section is completed, the arrangement of the persons is required to be carried out and analyzed after data is obtained, 3 to 4 persons are required to spend 1d time when 1 shield section is obtained, 20 sections of 1 typical subway tunnel line are taken as an example, the time is required to be about 800h and 160 persons when the detection of 1 line is completed, in the detection process, the manual detection efficiency is reduced due to the neglect of detection points and errors of data caused by negligence of workers, meanwhile, the detection and analysis period of the traditional method is too long and the person requirement is too large, and the detection efficiency of the waterproof performance of the subway shield tunnel is greatly reduced. Therefore, a more effective shield tunnel waterproof performance monitoring system is needed to meet the increasingly huge subway system waterproof monitoring requirement.
With the rapid development of artificial intelligence in various industries, the detection precision of the intelligent monitoring system is continuously improved particularly in the field of mechanical manufacturing, and an artificial detection method is replaced to a great extent, so that the system is also suitable for a subway shield tunnel monitoring system.
Disclosure of Invention
The invention aims to provide a subway shield tunnel segment seam waterproof performance monitoring system, and provides an intelligent monitoring method capable of accurately predicting and early warning shield tunnel seam waterproof capacity on the basis of the subway shield tunnel segment seam waterproof performance monitoring system. The technical scheme is as follows:
an early warning type intelligent monitoring method for the waterproof performance of a segment joint of a subway shield tunnel is used for monitoring the waterproof performance of the segment joint, an adopted segment joint monitoring system comprises a displacement sensor monitoring device, a slide rail axially fixed on the inner wall of a tunnel segment, a movable data receiver and an upper computer, wherein,
the displacement sensor monitoring device comprises an optical fiber displacement sensor, a reflector plate, a laser light source, a photoelectric converter and a data acquisition unit;
the optical fiber displacement sensor comprises an input optical fiber and an output optical fiber, and the input optical fiber and the output optical fiber are pre-embedded on the side surface of a segment joint to conduct laser;
the reflecting sheet comprises a plurality of groups of reflecting strips with different specular reflectivities, and is arranged on the opposite joint surface with the joint surface of the optical fiber sensor, and the reflecting belts with different specular reflectivities perform multi-stage feedback on the input light intensity;
the laser light source is connected with the input optical fiber and provides laser light intensity for the input optical fiber;
the photoelectric converter is connected with the output optical fiber and converts the optical signal into an electrical signal;
the data acquisition unit is connected with the photoelectric converter and used for storing shield tunnel monitoring real-time data;
the mobile data receiver can move along the sliding rail and is used for regularly receiving and conveying data of the data collectors distributed in the shield tunnel to an upper computer.
The monitoring method realized by adopting the system comprises the following steps: the upper computer performs machine learning based on a BP neural network algorithm: training a plurality of segment joint waterproof water tightness test data samples, reducing errors of a target output value and an actual output value according to a gradient descending feedforward transmission and back propagation learning rule, and establishing a characteristic relation between joint opening amount and wrong station amount and waterproof capacity, so that decision and early warning are carried out on operation period tunnel segment joint monitoring data by utilizing a neural network, and monitoring is carried out according to the following steps:
(1) data sample collection and processing
The collected data are mainly obtained from a plurality of groups of tests of water tightness of the segment joints, and the quantity S of the collected data at different staggered platforms is obtainediWater-proof property of joint under condition Paver,waterOpening amount delta from jointiThe non-linear relationship between the two, each set of test data includes the following parameters: opening amount (delta) of segment jointsi) Segment joint dislocation amount (S)i) Hardness of sealing gasket (A), waterproof ability of joint (P)aver,water) Grouping the data into training data and test data;
normalizing the acquired data to convert the acquired data into corresponding values in a [0,1] interval;
(2) establishment of BP neural network
The method comprises the following steps of establishing a shield segment joint waterproof performance learning model based on a BP neural network, wherein the neural network model comprises an input layer, a hidden layer and an output layer, and the input layer comprises 3 neurons: the opening amount of the segment joints, the dislocation amount of the segment joints and the hardness of the sealing gasket; the output layer comprises a nerve unit, namely the waterproof capability of a segment joint; the neurons of the input layer and the output layer adopt the test sample data normalized in the step (1);
(3) neural network feedforward transfer and testing
Setting a neural network training stopping condition: inputting the test sample data subjected to normalization processing in the step (1) into the neural network learning model constructed in the step (2) to obtain the maximum iteration times and the training target error, substituting the segment seam waterproof test data into the training neuron matrix for inspection after the sample data training is finished, and calculating the training target error by adopting the mean square error;
(4) the output data of the neural network in the test data, namely the waterproof capability value, is transmitted reversely;
(5) and processing the segment joint monitoring data during operation by using the established neural network model.
Preferably, according to the nonlinear characteristics of input training samples, the activation functions of neural units of the hidden layer and the output layer in the neural network are derived by adopting Sigmoid type transformation functions; the neural network learning model is realized by a scimit-learn library in Python. A
(5) The method comprises the following steps: opening delta for monitoring tunnel segment joints by using neural network modeli,monitorSum of station quantity Si,monitorDiagnosing and analyzing, outputting the actual value of the waterproof capacity of the joint, and judging and analyzing:
in the formula, Pneuron,waterOutputting seam waterproof ability value, P, for neural network modeldesign,waterA design value for waterproofing a segment joint;
when the conditions meet the first condition, the waterproof capacity of the segment joints is normal; and when the condition meets the second condition, the waterproof capability of the segment joints is invalid, the waterproof failure reasons of the tunnel are diagnosed, the change trend of the waterproof capability of the joints is predicted and early warning is carried out by positioning and counting the distribution and monitoring values of the waterproof failure points of the joints.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. according to the invention, the reflective displacement sensor is arranged at the shield segment joint, so that the opening amount and the slab staggering amount of the segment joint can be monitored in real time based on the light intensity change of the input optical fiber and the output optical fiber;
2. according to the segment joint monitoring system, the displacement sensor is arranged at the joint and the data collector is arranged at the segment, so that unmanned monitoring of the segment joint of the shield tunnel is realized, and the monitoring precision and efficiency are effectively improved;
3. the intelligent learning and decision-making method for the waterproof capability of the joints is based on a BP neural network algorithm to complete machine learning, so that a neural network model with waterproof capability evaluation capability is formed, the neuron model diagnoses and analyzes segment joint monitoring data, the intelligent evaluation of the waterproof capability of the joints is realized, and the service life of a subway shield tunnel is effectively prolonged;
drawings
FIG. 1 is a flow chart of intelligent monitoring and early warning of waterproof capability of a segment joint of a subway shield tunnel according to the invention;
FIG. 2 is a schematic view of a subway shield tunnel segment seam monitoring system according to the present invention;
FIG. 3 is a schematic diagram of a reflective displacement sensor according to the present invention;
FIG. 4 is a schematic layout view of a reflective displacement sensor according to the present invention;
FIG. 5 is a simulation graph of an optical fiber micro-displacement sensor with reflected light intensity under different reflectivity conditions;
FIG. 6 is a diagram of a neural network learning model according to the present invention.
FIGS. 1 to 6 include:
1-subway shield tunnels;
2-tunnel vault;
3-tunnel arch waist;
4, tunnel arch bottom;
5, longitudinal seam of the tunnel;
6, circularly sewing the tunnel;
7-a data collector;
8-a slide rail;
9-a mobile data receiver;
10-an optical fiber displacement sensor;
11-input fiber;
12-an output fiber;
13-self-adhesive reflective sheet;
14-a reflective strip;
15-rubber gasket;
16-outside the sealing groove;
17-inside the seal groove;
18-a laser light source;
19-photoelectric converter.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings:
the invention provides an early warning type subway shield tunnel segment seam waterproof performance intelligent monitoring system based on a BP artificial neural network intelligent algorithm aiming at the problems of a traditional subway shield tunnel seam waterproof detection method, a seam waterproof performance neural network model is formed by supervising and learning sample data which represents seam waterproof capacity according to seam opening amount and slab staggering amount, the monitoring opening amount and the slab staggering amount of a formed shield tunnel segment seam are diagnosed and analyzed, and the prediction and early warning of the segment seam waterproof capacity are made by taking the monitoring opening amount and the slab staggering amount as an intermediate bridge.
The intelligent monitoring system for the waterproof capability of the segment joints of the subway shield tunnel shown in the figure 1 consists of 3 parts, namely a segment joint monitoring system, an intelligent learning system for the waterproof capability of joints and a prediction and early warning system for the waterproof capability of joints.
The segment joint monitoring system comprises a displacement sensor monitoring device, a sliding rail, a movable data receiver and an upper computer, wherein the sliding rail is axially fixed on the inner wall of a tunnel segment, and the displacement sensor monitoring device comprises an optical fiber displacement sensor, a self-adhesive reflector, a laser light source, a photoelectric converter and a data acquisition unit.
Referring to fig. 2, the segment joints of the subway shield tunnel 1 are divided into a longitudinal tunnel joint 5 between segments and a circular tunnel joint 6 between a tube ring and a tube ring, the longitudinal tunnel joint 5 and the circular tunnel joint 6 are main waterproof parts of the shield tunnel, and according to existing researches, the tunnel vault 2, the tunnel arch 3 and the tunnel arch bottom 4 of the subway shield tunnel 1 are easy to leak during operation, and the factor generally causing seepage of the joints is the joint opening amount deltaiExcessive and wrong station quantity SiIf the size of the shield tunnel is too large, the opening amount and the slab staggering amount of the segment joints are closely related to the contact stress of the sealing gaskets, so that the monitoring system is mainly characterized in that reflective optical fiber displacement sensors (not shown in the figure) are arranged at the tunnel vault 2, the tunnel arch 3 and the tunnel arch bottom 4, the change of the opening amount and the slab staggering amount output in the time domain in the operation period is used for representing the change of the waterproof capability of the joints, the data collector 7 receives, stores and transmits the monitoring data of the reflective optical fiber displacement sensors at the segment joints, and the mobile data receiver 9 regularly receives the data of the data collector 7 distributed in the shield tunnel along the slide rail 8 and transmits the data to an upper computer.
Seam opening delta as described with reference to figure 3iAnd the amount of dislocation SiMonitoring devices mainly comprises optic fibre displacement sensor 10 and from pasting formula reflector plate 13, optic fibre displacement sensor 10 and from pasting formula reflector plate 13 and being located section of jurisdiction seam face both sides respectively, including input fiber 11 and output fiber 12 in optic fibre displacement sensor 10, light gets into in the input fiber 11 from the light source coupling, through being reflected to output fiber 12 again from pasting formula reflector plate 13, because input fiber 11 and output fiber 12's light intensity has difference to some extent, measure the section of jurisdiction seam based on this principle, set for 11 fine end exit optical field distributions of input fiber to axisymmetric gaussian distribution model, receive the light intensity to input fiber 11 outgoing light intensity and output fiber 12 and deduce and be:
in the formula, I (Delta)i) For the output fibre 12 to receive the light intensity, I0The intensity of the light emitted from the input optical fiber 11, a is the radius of the optical fiber, RiIs the specular reflectivity, Delta, of the self-adhesive reflector 13iOpening amount of segment joint thetacIs the angle of reflection.
The self-adhesive reflector 13 is composed of different groups of reflector strips 14, and the calculation parameters of each group of reflectors are respectively as follows from top to bottom:
[R-i,H-i],……,[R-1,H-1]，[R0,H0]，[R1,H1],……,[Ri,Hi]
wherein R isiThe upper and lower mirror reflectivities are different orders of magnitude respectivelyiThe width of the reflective strip 14 is H since the measurement gap is 1mm in magnitude-i＝……＝H-1＝H0＝……＝HiDifferent seam staggering values are set by different specular reflectances, 1 mm.
When the output fiber 12 is located within the reflected light cone, the reflected light is received by the output fiber 12 and is obtained at a different RiUnder the condition of I (. DELTA.)i)～ΔiThe relationship of (1). Dislocation quantity S between jointsiFrom specular reflectivities R of different orders of magnitudeiThe calculated curve characteristic is judged, and the opening quantity delta between joints is judgediFrom specular reflectivity RiI (. DELTA.) under the conditionsi)～ΔiThe curve can be derived.
Referring to fig. 4, the waterproof capability of the segment joint is mainly realized by compressing the rubber gasket 15, when the shield tunnel is in a deep stratum, the reflective optical fiber displacement sensors of the tunnel vault 2 and the tunnel vault 4 are mainly arranged on the inner side 17 of the seal groove, and the reflective optical fiber displacement sensor of the tunnel vault 3 is mainly arranged on the outer side 16 of the seal groove. When the shield tunnel is in a shallow stratum, the reflective optical fiber displacement sensors of the tunnel vault 2 and the tunnel vault bottom 4 are mainly arranged on the outer side 16 of the sealing groove, and the reflective optical fiber displacement sensors of the tunnel vault 3 are mainly arranged on the inner side 1 of the sealing groove7. The optical fiber displacement sensor 10 of the reflection type optical fiber displacement sensor is pre-buried in a duct piece, the fiber end is flush with the joint surface, the self-sticking type reflector plate 13 is stuck to the joint surface, the optical fibers of the optical fiber sensor 10 are arranged in parallel along the joint surface of the duct piece, the input optical fiber 11 is externally connected with a laser light source 18, the output optical fiber 12 is externally connected with a photoelectric converter 19, and optical signals pass through the photoelectric converter 19 to enable the light intensity and the opening amount delta to be deltaiIs converted into a voltage and an opening amount deltaiThe data collector 7 collects and stores the electrical signal output by the photoelectric converter 19.
The present invention is further described in detail by combining the embodiments with the reflective optical fiber displacement sensor. Using I (Delta)i) The expressed optical fiber output characteristic modulation function is expressed by using the following formula to express the optical fiber output light intensity influence parameter:
I(Δi)＝f(a,Ri,θc)
MATLAB software is used for controlling the numerical simulation of the variable method to simulate the influence of different mirror reflectivities on the optical fiber displacement sensor, and the theoretical curve of the optical fiber micro-displacement sensor with reflected light intensity under the condition of different reflectivities is obtained, as shown in figure 5. Wherein, the emergent light intensity I of the input optical fiber 110＝60×108cd, radius a of input fiber 11 and output fiber 12 is 0.2mm, and reflection angle θc15 DEG, self-adhesive reflection sheet 13 specular reflectance RiAccording to different dislocation quantities SiThe reflection bars 14 are calibrated to have 10 reflectivities of 0.1, 0.2, … …, 0.9, 1.0, etc.
The intelligent learning system for the seam waterproof capability shown in fig. 6 mainly performs machine learning based on a BP neural network algorithm, trains a large number of segment seam waterproof and watertight test data samples, reduces errors between a target output value and an actual output value according to a gradient descending feedforward transmission and back propagation learning rule, and effectively establishes a characteristic relation between a seam opening amount and a slab staggering amount and waterproof capability, so that decision and early warning can be performed on tunnel segment seam monitoring data in an operation period by using a neural network.
The algorithm learning mainly comprises the following steps:
(1) data sample collection and processing
The collected data are mainly obtained from a plurality of groups of tests of the water tightness of the segment joints, and the test contents are S at different dislocation quantitiesiWater-proof property of joint under condition Paver,waterOpening amount delta from jointiThe non-linear relationship between them, and therefore, each set of experimental data includes the following parameters: opening amount (delta) of segment jointsi) Segment joint dislocation amount (S)i) Hardness of sealing gasket (A), waterproof ability of joint (P)aver,water) The data is grouped into training data and test data.
Because the magnitude of each parameter is very different, in order to accelerate the convergence speed of the neural network, the collected data needs to be normalized to be converted into [0,1]]Corresponding value within the interval
The calculation formula is as follows:
in the formula, XminIs the minimum value of each parameter, XmaxIs the maximum value of each parameter, XiFor the data collected for each of the parameters,
and normalizing the processed values of the parameters.
(2) Establishment of BP neural network
The method comprises the following steps of establishing a shield segment joint waterproof performance learning model based on a BP neural network, wherein the neural network model comprises an input layer, a hidden layer and an output layer, and the input layer comprises 3 neurons: the opening amount of the segment joints, the dislocation amount of the segment joints and the hardness of the sealing gasket; the output layer comprises a nerve unit, namely the waterproof capability of a segment joint; the number of nodes of the hidden layer can be obtained by an empirical formula:
in the formula, m is the number of nodes of an input layer, n is the number of nodes of an output layer, and a is an adjusting constant between 1 and 10. The number of nodes of the hidden layer in the neural network is 6 according to calculation.
According to the nonlinear characteristics of input training samples, the activation functions of neural units of a hidden layer and an output layer in the neural network are derived by adopting Sigmoid type transformation functions, and the formula is as follows:
the neural network learning model can be realized by a scimit-learn library in Python, and the neurons of the input layer and the output layer adopt the test sample data subjected to normalization processing in the step (1).
(3) Neural network feedforward transfer and testing
Setting a neural network training stopping condition: and (3) inputting the test sample data subjected to normalization processing in the step (1) into the neural network learning model constructed in the step (2) for the maximum iteration times and the training target error, and substituting the segment seam waterproof test data into the training neuron matrix for inspection after the sample data training is finished. The training target error is calculated by adopting a mean square, and the formula is as follows:
in the formula, Q is the number of neuron training groups in the input layer, Y (k) is the predicted value of the neural network, and t (k) is the actual output value of the output layer of the neural network, so that the mean square error is smaller than the error of the training target.
(4) Back propagation of neural network output data, i.e., water resistance values, in test data
After the neural network test is finished, performing inverse normalization processing on output data, wherein the processing flow comprises the following steps:
wherein Y is the output value after inverse normalization, YminIs the minimum value of the waterproof capability of the pipe sheet joint in the water tightness test data, YmaxThe maximum value of the waterproof capability of the pipe sheet joint in the water tightness test data,
and outputting the actual value for the neural network.
And comparing the water tightness test data with the output value of the inverse normalization of the neural network model, and evaluating the accuracy and the applicability of the training neural network learning model.
(5) Processing segment joint monitoring data during operation by utilizing established neural network model
Opening amount delta for monitoring tunnel segment joints by using neural network model with applicabilityi,monitorSum of station quantity Si,monitorDiagnosing and analyzing, outputting the actual value of the waterproof capacity of the joint, and judging and analyzing:
in the formula, Pneuron,waterOutputting seam waterproof ability value, P, for neural network modeldesign,waterAnd designing the waterproof design value of the segment joint.
When the conditions meet the first condition, the waterproof capacity of the segment joints is normal; and when the condition is satisfied, the waterproof capability of the segment joints is invalid, the waterproof failure reasons of the tunnel are diagnosed and the change trend of the waterproof capability of the joints is predicted by positioning and counting the distribution and monitoring values of the waterproof failure points of the joints, the analysis and prediction are early warned, and related treatment suggestions are provided.
The above description is only a preferred embodiment of the present invention, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be understood that any modification, equivalent replacement, or improvement made without departing from the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (3)
1. An early warning type intelligent monitoring method for the waterproof performance of a segment joint of a subway shield tunnel is used for monitoring the waterproof performance of the segment joint, an adopted segment joint monitoring system comprises a displacement sensor monitoring device, a slide rail axially fixed on the inner wall of a tunnel segment, a movable data receiver and an upper computer, wherein,
the displacement sensor monitoring device comprises an optical fiber displacement sensor, a reflector plate, a laser light source, a photoelectric converter and a data acquisition unit;
the optical fiber displacement sensor comprises an input optical fiber and an output optical fiber, and the input optical fiber and the output optical fiber are pre-embedded on the side surface of a segment joint to conduct laser;
the reflecting sheet comprises a plurality of groups of reflecting strips with different specular reflectivities, the reflecting strips are arranged on the opposite joint surfaces with the joint surfaces of the optical fiber sensor, and the reflecting strips with different specular reflectivities perform multi-stage feedback on the input light intensity;
the laser light source is connected with the input optical fiber and provides laser light intensity for the input optical fiber;
the photoelectric converter is connected with the output optical fiber and converts the optical signal into an electrical signal;
the data acquisition unit is connected with the photoelectric converter and used for storing shield tunnel monitoring real-time data;
the mobile data receiver can move along the sliding rail and is used for receiving data of the data collector distributed in the shield tunnel and transmitting the data to the upper computer;
the monitoring method realized by adopting the system comprises the following steps: the upper computer performs machine learning based on an algorithm: training a plurality of segment joint waterproof water tightness test data samples, reducing errors of a target output value and an actual output value according to a gradient descending feedforward transmission and back propagation learning rule, and establishing a characteristic relation between joint opening amount and wrong station amount and waterproof capacity, so that decision and early warning are carried out on operation period tunnel segment joint monitoring data by utilizing a neural network, and monitoring is carried out according to the following steps:
(1) data sample collection and processing
The collected data are mainly obtained from a plurality of groups of tests of water tightness of the segment joints, and the quantity S of the collected data at different staggered platforms is obtainediWater-proof property of joint under condition Paver,waterOpening amount delta from jointiThe non-linear relationship between the two, each set of test data includes the following parameters: opening amount delta of segment jointsiSegment joint dislocation amount SiHardness A of sealing gasket and waterproof capability P of jointaver,waterGrouping the data into training data and test data;
normalizing the acquired data to convert the acquired data into corresponding values in a [0,1] interval;
(2) establishment of BP neural network
The method comprises the following steps of establishing a shield segment joint waterproof performance learning model based on a BP neural network, wherein the shield segment joint waterproof performance learning model comprises an input layer, a hidden layer and an output layer, wherein the input layer comprises 3 neurons: the opening amount of the segment joints, the dislocation amount of the segment joints and the hardness of the sealing gasket; the output layer comprises a nerve unit, namely the waterproof capability of a segment joint; the neurons of the input layer and the output layer adopt the test sample data normalized in the step (1);
(3) neural network feedforward transfer and testing
Setting a BP neural network training stopping condition: inputting the test sample data subjected to normalization processing in the step (1) into the shield segment seam waterproof performance learning model constructed in the step (2), substituting segment seam waterproof test data into a training neuron matrix for inspection after sample data training is completed, and calculating a training target error by adopting a mean square error;
(4) BP neural network output data in the test data, namely waterproof capability values are transmitted reversely;
(5) and processing the segment joint monitoring data during operation by using the established shield segment joint waterproof performance learning model.
2. The method of claim 1, wherein according to the nonlinear characteristics of input training samples, the activation functions of neural units of the hidden layer and the output layer in the shield segment joint waterproof performance learning model are derived by using Sigmoid type transformation functions; the shield segment joint waterproof performance learning model is realized through a scimit-learn library in Python.
3. The method of claim 1, wherein (5) is performed as follows: opening amount delta for monitoring tunnel segment joints by utilizing shield segment joint waterproof capability learning modeli,monitorSum of station quantity Si,monitorDiagnosing and analyzing, outputting the actual value of the waterproof capacity of the joint, and judging and analyzing:
in the formula, Pneuron,waterActual value of joint waterproof capability, P, output by the model for learning waterproof capability of the joints of the shield segmentsdesign,waterA design value for waterproofing a segment joint;
when the conditions meet the first condition, the waterproof capacity of the segment joints is normal; and when the condition meets the second condition, the waterproof capability of the segment joints is invalid, the waterproof failure reasons of the tunnel are diagnosed, the change trend of the waterproof capability of the joints is predicted and early warning is carried out by positioning and counting the distribution and monitoring values of the waterproof failure points of the joints.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910209178.XA CN110006348B (en) | 2019-03-19 | 2019-03-19 | Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910209178.XA CN110006348B (en) | 2019-03-19 | 2019-03-19 | Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110006348A CN110006348A (en) | 2019-07-12 |
| CN110006348B true CN110006348B (en) | 2021-05-11 |
Family
ID=67167398
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910209178.XA Active CN110006348B (en) | 2019-03-19 | 2019-03-19 | Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110006348B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112324457B (en) * | 2020-10-16 | 2021-08-03 | 中铁十四局集团有限公司 | Water-rich super-large section shield tunnel waterproof composite lining structure and construction method |
| CN112324456B (en) * | 2020-10-16 | 2022-02-08 | 中铁十四局集团有限公司 | Shield tunnel composite lining structure and method |
| CN113569362B (en) * | 2021-09-22 | 2021-12-21 | 南昌市建筑科学研究所（南昌市建筑工程质量检测中心） | System and method for managing whole life cycle of shield segment |
| CN114518203B (en) * | 2022-01-26 | 2024-02-20 | 中交隧道工程局有限公司 | Shield tunnel segment leakage point identification method |
| CN114858079B (en) * | 2022-05-23 | 2022-11-25 | 青岛益群地下城开发有限公司 | Tunnel deformation monitoring method based on distributed multipoint strain and displacement conversion network |
| CN115711706B (en) * | 2023-01-09 | 2023-04-18 | 中南大学 | Tunnel joint sealing gasket waterproof capacity early warning system and judgment method |
| CN116989680B (en) * | 2023-09-28 | 2023-12-29 | 武汉科技大学 | Fiber bragg grating monitoring device and method for joints in shield segment assembly process |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3240417B2 (en) * | 1992-08-25 | 2001-12-17 | 清水建設株式会社 | Construction method of headrace tunnel and headrace tunnel |
| CN202690083U (en) * | 2012-01-13 | 2013-01-23 | 上海市政工程设计研究总院（集团）有限公司 | Waterproof structure of shield tunnel duct piece |
| US8973445B2 (en) * | 2012-09-05 | 2015-03-10 | Qorex Llc | Multi-mode holographic pressure sensor |
| CN104240470A (en) * | 2013-06-09 | 2014-12-24 | 同济大学 | Wireless joints displacement measuring system for metro shield tunnel |
| CN105987778B (en) * | 2015-05-06 | 2018-10-30 | 中铁建设投资集团有限公司 | A kind of in-situ measuring method of duct pieces of shield tunnel seam stress |
| CN207033473U (en) * | 2017-07-17 | 2018-02-23 | 中铁隧道勘测设计院有限公司 | A kind of waterproof structure of shield tunnel segment seam |
| CN108489456A (en) * | 2018-03-30 | 2018-09-04 | 胡仲春 | tunnel monitoring system and tunnel monitoring method |
| CN109387385B (en) * | 2018-10-24 | 2019-09-03 | 中国科学院武汉岩土力学研究所 | A kind of three-dimensional TBM shield outer surface country rock Load Identification Methods |
-
2019
- 2019-03-19 CN CN201910209178.XA patent/CN110006348B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110006348A (en) | 2019-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110006348B (en) | Intelligent monitoring method for waterproof performance of segment joint of early warning type subway shield tunnel | |
| CN110044268B (en) | Shield tunnel joint opening and slab staggering monitoring system based on optical fiber reflection principle | |
| US8306276B2 (en) | Bridge structural safety monitoring system and method thereof | |
| RU2563419C2 (en) | Method of monitoring of technical state of pipeline and system for its implementation | |
| CN106156336A (en) | A kind of Cable-Stayed Bridge Structure state evaluation system and assessment method | |
| CN109323821B (en) | Multi-station gas detector verification instrument and verification method | |
| KR102163910B1 (en) | A system for predicting tension of mooring lines through location information prediction and location information learning based on ocean information of the offshore platform | |
| CN110672154A (en) | Civil engineering building monitoring system | |
| CN114169548A (en) | BIM-based highway bridge management and maintenance PHM system and method | |
| CN103994333A (en) | Oil gas pipe network leak detection method based on two-dimensional information fusion | |
| CN115329812A (en) | Road infrastructure abnormity monitoring method based on artificial intelligence | |
| CN111628570A (en) | Hydropower station safety monitoring fault diagnosis method and system | |
| CN117371337B (en) | Water conservancy model construction method and system based on digital twin | |
| CN107782785B (en) | Underground pipeline safety assessment method | |
| CN108776040B (en) | Bridge safety inspection system and diagnosis method | |
| CN116840135A (en) | Steel gate accelerated degradation test bed with simultaneous effects of corrosion and fatigue and post-degradation running state evaluation method | |
| CN116519054A (en) | Health state monitoring system and method for heat station equipment | |
| CN116337018A (en) | Water resource on-line monitoring system | |
| CN116305699B (en) | Pipeline supervision system based on omnibearing sensing | |
| CN116754022A (en) | On-line detection emergency early warning method and system for cable tunnel | |
| Pang | Intelligent belt conveyor monitoring and control | |
| Yu et al. | The design and evaluation of tunnel health monitoring system | |
| CN113567247A (en) | Bridge detection information management system | |
| Omar | Condition Assessment of Concrete Bridge Decks Using Ground and Airborne Infrared Thermography | |
| CN117636073B (en) | Concrete defect detection method, device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |