CN111609805A - Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature - Google Patents
Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature Download PDFInfo
- Publication number
- CN111609805A CN111609805A CN202010329370.5A CN202010329370A CN111609805A CN 111609805 A CN111609805 A CN 111609805A CN 202010329370 A CN202010329370 A CN 202010329370A CN 111609805 A CN111609805 A CN 111609805A
- Authority
- CN
- China
- Prior art keywords
- tunnel structure
- curvature
- tunnel
- strain
- section
- 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.)
- Granted
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
- G01B11/16—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge
- G01B11/18—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge using photoelastic elements
-
- 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
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
- G01B11/255—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures for measuring radius of curvature
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (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 belongs to the field of tunnel structure damage diagnosis, and particularly relates to a tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature. According to the method, a conversion relation model of the strain and curvature of the measuring point of the tunnel structure is constructed, a method for judging the accuracy of the conversion relation model of the strain and curvature of the measuring point of the tunnel structure is provided, the characteristics that the curvature can reflect the integral deformation of the structure in the influence range are utilized, the influence area of the strain of the measuring point of the structure is expanded, the sensitivity of the strain on the structure is improved, and the problem that the capability of processing high-measuring-point-density strain data by using the conventional damage diagnosis method is low is solved; on the basis, a tunnel structure damage diagnosis factor based on multi-section curvature is constructed by using a BP artificial neural network, so that the tunnel structure damage is diagnosed.
Description
Technical Field
The invention belongs to the field of tunnel structure damage diagnosis, and particularly relates to a tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature.
Background
The structural health monitoring technology can continuously and reliably provide tunnel structure response and environmental load information, identify structural defects in design and construction, diagnose and evaluate structural damage in operation, so the technology is widely applied to tunnel engineering. The following problems still exist in the process of utilizing monitoring and detection technology to guarantee the operation safety of the tunnel structure: firstly, the conventional detection technology mainly detects the information of vault sinking, peripheral convergence displacement, surrounding rock, lining pressure and the like of the key section of the tunnel structure regularly or irregularly, and cannot realize the real-time diagnosis of the state of the tunnel structure; secondly, the existing tunnel structure health monitoring mainly adopts a point type sensor technology, measuring points of the sensor are distributed discontinuously in space, networking is difficult, the possibility of missing measurement of key parts of the structure exists, and accurate diagnosis of tunnel structure damage is difficult to realize. Thirdly, the existing method for diagnosing damage based on structural strain response can only reflect the local information of the structure, has a small identification area sensitive to the damage of the tunnel structure, and has weak capability of processing high-measuring-point density strain data. In order to solve the problems, the strain response information of the tunnel structure, which is acquired by a monitoring technology with large range, long distance and long duration, is required to be utilized to construct a physical quantity with higher sensitivity to the damage of the tunnel structure, and a corresponding damage diagnosis method is provided.
Disclosure of Invention
The invention aims to provide a tunnel structure state diagnosis method based on full-distribution strain measuring point section curvature for realizing tunnel structure damage diagnosis. The method specifically comprises the following steps:
the method comprises the following steps: a conversion relation model of the strain of the measuring point of the tunnel structure and the curvature is constructed by using the strain monitoring data of the tunnel structure acquired by the full-distribution optical fiber sensors, and based on the characteristic that the curvature can reflect the integral deformation of the structure in the influence range, the influence area of the strain of the measuring point of the structure is expanded, and the sensitivity of the strain to the structural damage is improved;
step two: analyzing the influence of the shearing force and the bending moment on a conversion relation model of the strain and the curvature of the measuring point of the tunnel structure, and judging the accuracy of the conversion relation model of the strain and the curvature of the measuring point of the tunnel structure;
step three: constructing a tunnel structure damage diagnosis factor based on multi-section curvature by using the structure monitoring curvature obtained by converting the tunnel structure strain monitoring data through the conversion relation model in the step one and combining a BP artificial neural network;
step four: and under the health state of the tunnel structure, calculating a tunnel structure damage diagnosis threshold value. On the basis, the accumulated damage diagnosis factor exceeding the threshold value is calculated, and damage diagnosis decision is carried out on the tunnel structure.
The invention has the beneficial effects that:
the tunnel structure state diagnosis method based on the full-distribution strain measuring point section curvature constructs a conversion relation model of the strain and the curvature of the structure measuring point, constructs a tunnel structure damage diagnosis factor based on the multi-section curvature by utilizing a BP artificial neural network, and realizes the diagnosis of the tunnel structure damage. The method utilizes the characteristic that the curvature can reflect the integral deformation of the structure within the influence range, expands the influence area of the strain of the measuring point of the structure, improves the sensitivity of the strain on the structure, solves the problem that the existing damage diagnosis method has low capability of processing high measuring point density strain data, has good robustness, and is easier to diagnose the potential damage of the tunnel structure and has higher anti-noise capability compared with the traditional method.
Drawings
FIG. 1 is a flowchart of a tunnel structural state diagnosis method based on a fully distributed strain point cross-sectional curvature according to an embodiment.
FIG. 2 is a diagram of a finite element model of a tunnel structure and surrounding rocks in example 1.
Fig. 3 is the accuracy of the diagnostic results at 3% noise level in example 1.
Fig. 4 shows the accuracy of the diagnosis result at 5% noise level in example 1.
Detailed Description
The first embodiment is as follows: the tunnel structure state diagnosis method based on the full-distribution strain measuring point section curvature is shown in figure 1, and comprises the following steps:
the method comprises the following steps: a conversion relation model of the strain of the measuring point of the tunnel structure and the curvature is constructed by using the strain monitoring data of the tunnel structure acquired by the full-distribution optical fiber sensors, and based on the characteristic that the curvature can reflect the integral deformation of the structure in the influence range, the influence area of the strain of the measuring point of the structure is expanded, and the sensitivity of the strain to the structural damage is improved;
step two: analyzing the influence of the shearing force and the bending moment on a relation model of the conversion of the strain and the curvature of the measuring point of the tunnel structure, and judging the accuracy of the conversion relation model of the strain and the curvature of the measuring point of the tunnel structure;
step three: constructing a tunnel structure damage diagnosis factor based on multi-section curvature by using the structure monitoring curvature obtained by converting the tunnel structure strain monitoring data through the conversion relation model in the step one and combining a BP artificial neural network;
step four: under the health state of the tunnel structure, calculating a tunnel structure damage diagnosis threshold value, and on the basis, performing damage diagnosis decision on the tunnel structure by calculating accumulated damage diagnosis factors exceeding the threshold value;
in the embodiment, a tunnel structure state diagnosis method based on the curvature of the cross section of the fully distributed strain measuring points is provided for solving the problem of how to improve the capacity of processing centimeter-level high-measuring-point density strain data by an algorithm and improve the sensitivity degree of damage to the tunnel structure. Firstly, constructing a conversion relation model of the strain and curvature of the measuring point of the structure, and expanding the influence area of the strain of the measuring point of the structure and improving the sensitivity of the strain to the damage of the structure by utilizing the characteristic that the curvature can reflect the integral deformation of the structure in the influence range; secondly, constructing a tunnel structure damage diagnosis factor based on multi-section curvature by using a BP artificial neural network; and finally, diagnosing the damage of the tunnel structure. The method has good robustness and noise resistance, and can accurately identify the tiny damage of the tunnel structure.
The second embodiment is as follows: the embodiment is further described with respect to a tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature in the first embodiment, and in the first embodiment, a method for constructing a conversion relationship model between the strain and curvature of a tunnel structure measurement point in the first step is as follows:
by utilizing strain measuring points acquired along the tunnel trend, the structural curvature of a micro-section at the X-shaped section of the tunnel section is calculated by adopting the following formula:
in the formula, rho is the curvature of the section X obtained by utilizing strain calculation;tin order to obtain the strain of the upper surface of the micro-segment,bis the micro-segment lower surface strain; h istAnd hbRespectively the distances from the upper edge and the lower edge of the section X to the neutral axis.
The third concrete implementation mode: in the present embodiment, the method for diagnosing a tunnel structure state based on a cross-sectional curvature of a fully distributed strain measurement point is further described, and in the present embodiment, the method for determining the accuracy of the model of the conversion relationship between the strain and the curvature of the tunnel structure measurement point in the second step is as follows:
step two, firstly: for the micro-segment at the X-shaped section of the tunnel, under the action of concentrated load of the span of the micro-segment, the structure generates shear bending, the shear energy of the micro-segment is calculated as,
wherein κ is the shear energy of the microbodies, A is the area of the cross section, b is the width of the cross section, IzIs the section moment of inertia, S is the area moment of the section;
step two: integrating kappa along the longitudinal direction of the tunnel, calculating to obtain the shearing energy of the whole structure,
in the formula uQThe shear energy of the whole structure is shown, G is the shear modulus, Q is the shear force of a section X, L is the total length of the structure, and X is a micro-section along the direction of the tunnel;
step two and step three: according to the unit force method, a unit force is applied to the middle section of the tunnel structure, the work of the unit force is equal to the sum of the bending energy and the shearing energy of the whole structure, and the following relationship can be obtained,
wherein, M (x) and Q (x) are respectively bending moment and shearing force at any section of the tunnel structure; Δ is the deformation per unit force, E is the modulus of elasticity; m '(x) and Q' (x) are the first derivatives of M (x) and Q (x), respectively;
step two, four: the ratio gamma of the integral shearing energy to the bending energy of the structure is obtained from the steps,
in the formula, alpha is a characteristic coefficient and represents the relative rigidity between the tunnel and the surrounding rock soil body.
Step two and step five: for an actual tunnel, whether the tunnel structure can be regarded as meeting the assumption of a flat section can be judged according to the numerical value of gamma, the upper limit value of gamma is 0.01, namely when the gamma is less than 0.01, the integral shearing energy of the structure is less than 1% of the bending energy, the influence of the integral shearing energy can be ignored, and the conversion relation model of the strain and the curvature of a measuring point is considered to be accurate.
The fourth concrete implementation mode: in this embodiment, the tunnel structure state diagnosis method based on the cross-sectional curvature of the fully distributed strain measurement points is further described, in this embodiment, the tunnel structure damage diagnosis factor construction method based on the multi-cross-sectional curvature in the third step is as follows:
step three, firstly: acquiring tunnel structure strain monitoring data of N longitudinal sections of the tunnel by using a fully distributed optical fiber sensor to obtain a tunnel structure section curvature sample set rho,
ρ=([ρ1ρ2…ρgρg+1…ρN]T)N×r(g∈(1,2,…,N)) (6)
in the formula, ρgIs the vector of the monitoring sample of the curvature of the g-th section, which is defined as follows,
ρg={ρg1,ρg2,…,ρgr}1×r(7)
in the formula, rho is section curvature calculated by using a conversion relation model of strain and curvature of a measuring point of the tunnel structure, and r is a sample rhogNumber of sampling points.
Step three: in order to establish the relationship between the curvatures of all longitudinal sections of the tunnel structure by using a BP artificial neural network, a sample set rho is divided in the following form,
in the formula, ρinputInput layer sample set, rho, for a BP artificial neural networkoutputAn output layer sample set of the BP artificial neural network;
step three: substituting the input layer and output layer sample sets into a BP artificial neural network for training, establishing a relation model between the input layer and output layer sample sets as shown in the following formula,
ρinput=L(ρoutput)+ (9)
wherein, L (-) is a functional relationship model between the input layer and the output layer of the BP artificial neural network, and is a residual matrix of the BP artificial neural network in a reference state, as shown in the following formula:
=[1 2…z…r](N-g)×r(10)
step three and four: the residual matrix is calculated using the following equation,
in the formula (I), the compound is shown in the specification,for the output layer sample set rho obtained by using the trained BP artificial neural networkoutputIs calculated according to the following formula,
step three and five: constructing a tunnel structure damage diagnosis factor by utilizing a residual matrix under a reference state obtained by calculating monitoring data of each section curvature under a healthy state of the tunnel structure,
d={d1,d2,…,dz,…,dr}(z∈(1,2,…,r)) (13)
in the formula (d)zThe diagnostic factor of the tunnel structure damage at the z-th moment in the reference state is defined as the following formula,
in the formula (I), the compound is shown in the specification,zfor the z-th column of the residual matrix in the reference state, μ and Θ are the mean vector and covariance matrix of the residual matrix, respectively, which are defined as follows.
Step three and six: for the state to be diagnosed and diagnosed, the monitoring data of the curvature of each section of the tunnel is utilized to obtain a monitoring curvature sample set rho' of N sections along the longitudinal direction of the tunnel, which is expressed by the following formula,
ρ′=([ρ′1ρ′2…ρ′gρ′g+1…ρ′N]T)N×v(g∈(1,2,…,N)) (17)
in the formula, v is the number of sampling points for monitoring the strain of the cross section of the tunnel structure in the state to be diagnosed. Obtaining a residual error matrix' in a state to be diagnosed by using the BP artificial neural network trained in the reference state,
in the formula is ρ'outputAn output layer sample set of the BP artificial neural network in a state to be diagnosed;the estimation matrix of the output layer sample set in the state to be diagnosed is obtained by utilizing the trained BP artificial neural network. On the basis, calculating a tunnel structure damage diagnosis factor vector d' under the state to be diagnosed,
d′={d1′,d2′,…,d′w,…,d′v}(w∈(1,2,…,v)) (19)
of formula (II) to'wIs the w-th element of d'wIs the w-th column of the residual matrix' in the state to be diagnosed.
The fifth concrete implementation mode: in this embodiment, the tunnel structure state diagnosis method based on the cross-sectional curvature of the fully distributed strain measurement points is further described in the first embodiment, and in this embodiment, the tunnel structure damage diagnosis decision method in the fourth step is as follows:
step four, firstly: under the health state of the tunnel structure, calculating a threshold value for diagnosing the damage of the tunnel structure, as shown in the following formula,
λ=[d1,d2,…,dr]0.95(21)
wherein λ is the tunnel structural damage diagnosis threshold [. degree]0.95Operator representation takes 95% confidence probability value, drThe factor is a tunnel structure damage diagnosis factor at the r-th moment in the reference state.
Step four and step two: judging whether the tunnel structure has potential damage possibility or not by using the damage diagnosis threshold lambda established in the structural health state, namely
In the formula, TwIs a judgment index of the potential damage of the tunnel structure, w is any monitoring time of the state to be diagnosed, d'wThe tunnel structure damage diagnosis factor at the w moment in the state to be diagnosed. When T iswA value of 1 indicates that the tunnel structure may be damaged, when TwA value of 0 indicates that no damage is generated to the tunnel structure.
The following examples were used to demonstrate the beneficial effects of the present invention:
example 1:
in the embodiment, a finite element model of a certain actual tunnel structure is selected as an example, and the tunnel structure state diagnosis method based on the curvature of the section of the fully distributed strain measurement points is verified. The tunnel structure type is a small clear distance tunnel, and the values of the material calculation parameters of the tunnel structure are shown in table 1. The finite element model of the tunnel structure created using the finite element analysis software is shown in fig. 2. When a finite element model of the tunnel is established, solid45 solid units are used for simulating surrounding rocks, anchor rod reinforced areas and middle partition wall parts; the shell63 shell cells were used to simulate tunnel bracing (temporary bracing, inverted arch and preliminary bracing); the mesh200 unit is used for dividing a plane unit grid and is used when the plane unit grid is stretched into a three-dimensional model. The length of the tunnel segment is established to be 300 m.
TABLE 1 Tunnel Material parameter value-taking Table
And simulating a full-distribution sensing optical cable to acquire the strain of the vault and the side wall of the tunnel structure, wherein the measuring point resolution of the sensing optical cable is 20 cm. The damage section P is 150m away from the starting point of the tunnel structure; meanwhile, the damaged section P is located in the middle of the two monitoring sections, namely the longitudinal distance between the section and the two monitoring sections which are adjacent to each other at the left and right is 10 cm. The tunnel structure damage is simulated by multiplying the stiffness of all the units in the area by a reduction coefficient of 5%. And converting the monitored strain data into curvature, and calculating a gamma value of 0.002, thereby indicating that the conversion relation model of the strain and the curvature of the measuring point is accurate.
Noise with different degrees is applied to the calculated tunnel curvature values to simulate the change of the tunnel curvature values in a time domain, curvature data of each monitoring section is substituted into the diagnosis method provided by the invention to diagnose the damage of the tunnel structure, the reliability of the diagnosis result is tested, and the accuracy of the diagnosis result under the noise levels of 3 percent and 5 percent is shown in fig. 3 and 4.
From the lesion diagnostic map and data, it can be seen that: under the noise of 3%, the diagnosis accuracy is 98.0%, and the diagnosis results of the damaged section are all correct; at a 5% noise level, the diagnostic accuracy was 95.7%, but the lesion cross-section had begun to develop occasional diagnostic errors. Therefore, according to the result analysis of the numerical simulation, the diagnostic algorithm can obtain a better diagnostic result for the structural noise within 5 percent.
Claims (6)
1. The tunnel structure state diagnosis method based on the full-distribution strain measuring point section curvature is characterized by comprising the following steps of:
the method comprises the following steps: constructing a conversion relation model of the strain and curvature of a measuring point of the tunnel structure by using the strain monitoring data of the tunnel structure acquired by the full-distribution optical fiber sensors;
step two: analyzing the influence of the shearing force and the bending moment on a conversion relation model of the strain and the curvature of the measuring point of the tunnel structure, and judging the accuracy of the conversion relation model of the strain and the curvature of the measuring point of the tunnel structure;
step three: constructing a tunnel structure damage diagnosis factor based on multi-section curvature by using the structure monitoring curvature obtained by converting the tunnel structure strain monitoring data through the conversion relation model in the step one and combining a BP artificial neural network;
step four: and under the health state of the tunnel structure, calculating a tunnel structure damage diagnosis threshold value, and on the basis, performing damage diagnosis decision on the tunnel structure by calculating a damage diagnosis factor exceeding the threshold value.
2. The tunnel structure state diagnosis method based on the full distribution strain measuring point section curvature is characterized in that in the first step, a conversion relation model construction method of the tunnel structure measuring point strain and curvature comprises the following steps:
by utilizing strain measuring points acquired along the tunnel trend, the curvature of the micro-segment structure at the X position of the tunnel section is calculated by adopting the following formula:
in the formula, rho is the curvature of the section X obtained by utilizing strain calculation;tin order to obtain the strain of the upper surface of the micro-segment,bis the micro-segment lower surface strain; h istAnd hbRespectively the distances from the upper edge and the lower edge of the section X to the neutral axis.
3. The tunnel structure state diagnosis method based on the full distribution strain measuring point section curvature as claimed in claim 1, wherein the method for judging the accuracy of the conversion relation model of the tunnel structure measuring point strain and curvature in the second step is as follows:
step two, firstly: for the micro-segment at the X-shaped section of the tunnel, under the action of concentrated load of the span of the micro-segment, the structure generates shear bending, the shear energy of the micro-segment is calculated as,
wherein κ is the shear energy of the microbodies, A is the area of the cross section, b is the width of the cross section, IzIs the section moment of inertia, S is the area moment of the section;
step two: integrating kappa along the longitudinal direction of the tunnel, calculating to obtain the shearing energy of the whole structure,
in the formula uQThe shear energy of the whole structure is shown, G is the shear modulus, Q is the shear force of a section X, L is the total length of the structure, and X is a micro-section along the direction of the tunnel;
step two and step three: according to the unit force method, a unit force is applied to the middle section of the tunnel structure, the work of the unit force is equal to the sum of the bending energy and the shearing energy of the whole structure, and the following relationship can be obtained,
wherein, M (x) and Q (x) are respectively bending moment and shearing force at any section of the tunnel structure; Δ is the deformation in units of force, E is the modulus of elasticity, and M '(x) and Q' (x) are the first derivatives of M (x) and Q (x), respectively;
step two, four: the ratio gamma of the integral shearing energy to the bending energy of the structure is obtained from the steps,
in the formula, alpha is a characteristic coefficient and represents the relative rigidity between the tunnel and the surrounding rock soil body, and mu is a Poisson ratio;
step two and step five: for an actual tunnel, according to the value of γ, it can be determined whether the tunnel structure can be regarded as satisfying the assumption of a plane section.
4. The tunnel structure state diagnosis method based on the full-distribution strain measuring point section curvature is characterized in that the tunnel structure damage diagnosis factor construction method based on the multi-section curvature in the third step is as follows:
step three, firstly: acquiring tunnel structure strain monitoring data of N longitudinal sections of the tunnel by using a fully distributed optical fiber sensor to obtain a tunnel structure section curvature sample set rho,
ρ=([ρ1ρ2… ρgρg+1…ρN]T)N×r(g∈(1,2,…,N)) (6)
in the formula, ρgIs the vector of the monitoring sample of the curvature of the g-th section, which is defined as follows,
ρg={ρg1,ρg2,…,ρgr}1×r(7)
in the formula, rho is section curvature calculated by using a conversion relation model of strain and curvature of a measuring point of the tunnel structure, and r is a sample rhogThe number of sampling points;
step three: in order to establish the relationship between the curvatures of all longitudinal sections of the tunnel structure by using a BP artificial neural network, a sample set rho is divided in the following form,
in the formula, ρinputInput layer sample set, rho, for a BP artificial neural networkoutputAn output layer sample set of the BP artificial neural network;
step three: substituting the input layer and output layer sample sets into a BP artificial neural network for training, establishing a relation model between the input layer and output layer sample sets as shown in the following formula,
ρinput=L(ρoutput)+ (9)
wherein, L (-) is a functional relationship model between the input layer and the output layer of the BP artificial neural network, and is a residual matrix of the BP artificial neural network in a reference state, as shown in the following formula:
=[1 2…z…r](N-g)×r(10)
step three and four: the residual matrix is calculated using the following equation,
in the formula (I), the compound is shown in the specification,for the output layer sample set rho obtained by using the trained BP artificial neural networkoutputIs calculated according to the following formula,
step three and five: constructing a tunnel structure damage diagnosis factor by utilizing a residual matrix under a reference state obtained by calculating monitoring data of each section curvature under a healthy state of the tunnel structure,
d={d1,d2,…,dz,…,dr}(z∈(1,2,…,r)) (13)
in the formula (d)zThe diagnostic factor of the tunnel structure damage at the z-th moment in the reference state is defined as the following formula,
in the formula (I), the compound is shown in the specification,zfor the z-th column of the residual matrix in the reference state, μ and Θ are the mean vector and covariance matrix of the residual matrix, respectively, which are defined as follows:
step three and six: for the state to be diagnosed and diagnosed, the monitoring data of the curvature of each section of the tunnel is utilized to obtain a monitoring curvature sample set rho' of N sections along the longitudinal direction of the tunnel, which is expressed by the following formula,
ρ′=([ρ′1ρ′2… ρ′gρ′g+1… ρ′N]T)N×v(g∈(1,2,…,N)) (17)
in the formula, v is the number of sampling points for monitoring the section strain of the tunnel structure in the state to be diagnosed, a residual matrix' in the state to be diagnosed is obtained by using a BP artificial neural network trained in a reference state,
in formula (II), ρ'outputAn output layer sample set in a to-be-diagnosed state of the BP artificial neural network;calculating a tunnel structure damage diagnosis factor vector d' under the state to be diagnosed by utilizing an estimation matrix of an output layer sample set under the state to be diagnosed, which is obtained by the trained BP artificial neural network,
d′={d′1,d′2,…,d′w,…,d′v}(w∈(1,2,…,v)) (19)
of formula (II) to'wIs the w-th element of d'wIs the w-th column of the residual matrix' in the state to be diagnosed.
5. The tunnel structure state diagnosis method based on the full distribution strain measurement point section curvature according to claim 1, characterized in that the tunnel structure damage diagnosis decision method in the fourth step is as follows:
step four, firstly: under the health state of the tunnel structure, calculating a threshold value for diagnosing the damage of the tunnel structure, as shown in the following formula,
λ=[d1,d2,…,dr]0.95(21)
wherein λ is the tunnel structural damage diagnosis threshold [. degree]0.95Operator representation takes 95% confidence probability value, drThe factor is a tunnel structure damage diagnosis factor at the r moment in the reference state;
step four and step two: judging whether the tunnel structure has potential damage possibility or not by using the damage diagnosis threshold lambda established in the structural health state, namely
In the formula, TwIs a judgment index of the potential damage of the tunnel structure, w is any monitoring time of the state to be diagnosed, d'wThe factor for diagnosing the damage of the tunnel structure at the w-th moment in the state to be diagnosed is TwA value of 1 indicates that the tunnel structure may be damaged, when TwA value of 0 indicates that no damage is generated to the tunnel structure.
6. The tunnel structure state diagnosis method based on the full distribution strain point section curvature as claimed in claim 3, characterized in that: and taking 0.01 as the upper limit value of gamma, namely when the gamma is less than 0.01, considering that the conversion relation model of the strain and the curvature of the measuring point of the tunnel structure is accurate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010329370.5A CN111609805B (en) | 2020-04-23 | 2020-04-23 | Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010329370.5A CN111609805B (en) | 2020-04-23 | 2020-04-23 | Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111609805A true CN111609805A (en) | 2020-09-01 |
CN111609805B CN111609805B (en) | 2021-06-01 |
Family
ID=72196330
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010329370.5A Active CN111609805B (en) | 2020-04-23 | 2020-04-23 | Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111609805B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112487533A (en) * | 2020-11-30 | 2021-03-12 | 北京航空航天大学 | Full-section stress sensing method of tunnel structure health monitoring system |
CN113252450A (en) * | 2021-04-14 | 2021-08-13 | 浙江大学 | Method for evaluating maximum bending degree of bending and dumping rock mass |
CN113705721A (en) * | 2021-09-08 | 2021-11-26 | 哈尔滨工业大学 | Joint probability density function difference diagnosis method for bridge support group void diseases |
CN114623776A (en) * | 2022-05-16 | 2022-06-14 | 四川省公路规划勘察设计研究院有限公司 | Tunnel damage prediction method based on tunnel deformation monitoring |
CN114777661A (en) * | 2022-04-08 | 2022-07-22 | 哈尔滨工业大学 | Tunnel section convergence deformation quantitative calculation method based on high-density measuring point strain |
CN114858126A (en) * | 2022-03-04 | 2022-08-05 | 哈尔滨工业大学 | Tunnel longitudinal settlement deformation quantitative calculation method based on high-density measuring point strain |
CN114894109A (en) * | 2022-02-09 | 2022-08-12 | 重庆交通大学 | Tunnel strain information acquisition method and related device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5232547A (en) * | 1992-07-01 | 1993-08-03 | Motorola, Inc. | Simultaneously measuring thickness and composition of a film |
JP2011164024A (en) * | 2010-02-12 | 2011-08-25 | Shimizu Corp | Method and device of calculating amount of deflection of structure provided with optical fiber sensor |
KR101543368B1 (en) * | 2013-12-12 | 2015-08-11 | 한국건설기술연구원 | Hybrid sensor for structure and condition diagnosis system using it |
CN107843203A (en) * | 2017-10-13 | 2018-03-27 | 同济大学 | The implementation method of base pit engineering three-dimensional monitoring network based on distributed fiberoptic sensor |
CN109655007A (en) * | 2018-12-07 | 2019-04-19 | 贵州桥梁建设集团有限责任公司 | A kind of interior pipe inner concrete deformation monitoring method being perfused of grand bridge tubular arch |
CN110728089A (en) * | 2019-09-29 | 2020-01-24 | 济南城建集团有限公司 | Large-span bridge stay cable structure damage diagnosis method based on BOTDA technology |
CN110737977A (en) * | 2019-10-10 | 2020-01-31 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | tunnel surrounding rock deformation prediction method and prediction device |
-
2020
- 2020-04-23 CN CN202010329370.5A patent/CN111609805B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5232547A (en) * | 1992-07-01 | 1993-08-03 | Motorola, Inc. | Simultaneously measuring thickness and composition of a film |
JP2011164024A (en) * | 2010-02-12 | 2011-08-25 | Shimizu Corp | Method and device of calculating amount of deflection of structure provided with optical fiber sensor |
KR101543368B1 (en) * | 2013-12-12 | 2015-08-11 | 한국건설기술연구원 | Hybrid sensor for structure and condition diagnosis system using it |
CN107843203A (en) * | 2017-10-13 | 2018-03-27 | 同济大学 | The implementation method of base pit engineering three-dimensional monitoring network based on distributed fiberoptic sensor |
CN109655007A (en) * | 2018-12-07 | 2019-04-19 | 贵州桥梁建设集团有限责任公司 | A kind of interior pipe inner concrete deformation monitoring method being perfused of grand bridge tubular arch |
CN110728089A (en) * | 2019-09-29 | 2020-01-24 | 济南城建集团有限公司 | Large-span bridge stay cable structure damage diagnosis method based on BOTDA technology |
CN110737977A (en) * | 2019-10-10 | 2020-01-31 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | tunnel surrounding rock deformation prediction method and prediction device |
Non-Patent Citations (1)
Title |
---|
梅竹等: "钢筋混凝土结构的模型更新混合试验方法", 《工程科学与技术》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112487533A (en) * | 2020-11-30 | 2021-03-12 | 北京航空航天大学 | Full-section stress sensing method of tunnel structure health monitoring system |
AU2021273655B2 (en) * | 2020-11-30 | 2023-04-06 | Beihang University | Full-section force bearing sensing method of tunnel-structure health monitoring system |
CN113252450A (en) * | 2021-04-14 | 2021-08-13 | 浙江大学 | Method for evaluating maximum bending degree of bending and dumping rock mass |
CN113252450B (en) * | 2021-04-14 | 2022-05-06 | 浙江大学 | Method for evaluating maximum bending degree of bending and dumping rock mass |
CN113705721A (en) * | 2021-09-08 | 2021-11-26 | 哈尔滨工业大学 | Joint probability density function difference diagnosis method for bridge support group void diseases |
CN113705721B (en) * | 2021-09-08 | 2023-05-23 | 哈尔滨工业大学 | Joint probability density function difference diagnosis method for beam bridge support group void diseases |
CN114894109B (en) * | 2022-02-09 | 2023-10-17 | 重庆交通大学 | Tunnel strain information acquisition method and related device |
CN114894109A (en) * | 2022-02-09 | 2022-08-12 | 重庆交通大学 | Tunnel strain information acquisition method and related device |
CN114858126B (en) * | 2022-03-04 | 2022-11-29 | 哈尔滨工业大学 | Tunnel longitudinal settlement deformation quantitative calculation method based on high-density measuring point strain |
CN114858126A (en) * | 2022-03-04 | 2022-08-05 | 哈尔滨工业大学 | Tunnel longitudinal settlement deformation quantitative calculation method based on high-density measuring point strain |
CN114777661A (en) * | 2022-04-08 | 2022-07-22 | 哈尔滨工业大学 | Tunnel section convergence deformation quantitative calculation method based on high-density measuring point strain |
CN114777661B (en) * | 2022-04-08 | 2022-10-21 | 哈尔滨工业大学 | Tunnel section convergence deformation quantitative calculation method based on high-density measuring point strain |
CN114623776B (en) * | 2022-05-16 | 2022-07-22 | 四川省公路规划勘察设计研究院有限公司 | Tunnel deformation monitoring-based tunnel damage prediction method |
CN114623776A (en) * | 2022-05-16 | 2022-06-14 | 四川省公路规划勘察设计研究院有限公司 | Tunnel damage prediction method based on tunnel deformation monitoring |
Also Published As
Publication number | Publication date |
---|---|
CN111609805B (en) | 2021-06-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111609805B (en) | Tunnel structure state diagnosis method based on full-distribution strain measurement point section curvature | |
CN104457681B (en) | Girder structure dynamic deflection monitoring method based on strain mode | |
CN103116759B (en) | A kind of damnification recognition method being applicable to spatial mesh structure | |
CN107315874B (en) | Sensor layout method for simultaneously acquiring local deformation and overall modal information of structure | |
WO2019001022A1 (en) | Structural model estimation-based joint arrangement method for strain gauge and accelerometer | |
CN109002673B (en) | Bridge foundation scouring identification method based on vehicle braking impact effect | |
CN108280294B (en) | A kind of cable arch structure damage combined recognising method based on modal parameter | |
CN111259605A (en) | Soil landslide monitoring, early warning and evaluating method | |
CN109490072B (en) | Detection system for civil engineering building and detection method thereof | |
CN101504333B (en) | Progressive health monitoring method for cable structure median cord system based on strain monitoring | |
CN101915733B (en) | Frequency change-based structural damage degree assessment method | |
CN111324949B (en) | Engineering structure flexibility recognition method considering noise influence | |
CN104133960A (en) | Improved optimal arranging method of static sensors | |
CN108875178A (en) | For reducing the probabilistic sensor arrangement method of distinguishing structural mode | |
CN113267814A (en) | Method and device for measuring and calculating field shear wave velocity | |
CN110728089B (en) | Large-span bridge stay cable structure damage diagnosis method based on BOTDA technology | |
CN113420482B (en) | Segment load orthogonal numerical inversion method based on structural internal force monitoring value | |
CN112085354B (en) | Bridge foundation scouring diagnosis method based on vehicle-induced power response cross-correlation index system | |
KR101348817B1 (en) | Method and system for estimation of error or omitted measured values using artificial neural network | |
CN112883478B (en) | Steel structure displacement prediction method, device, terminal equipment and system | |
CN112329287B (en) | P-y curve Bayes learning method based on test pile monitoring data | |
Wang et al. | A method for quantitative damage identification in a high-piled wharf based on modal strain energy residual variability | |
CN113139228A (en) | Monitoring point arrangement optimization method for large-span foundation pit complex support system structure | |
CN102998367B (en) | Damage identification method based on virtual derivative structure | |
CN117235540A (en) | Sensor dynamic information linkage analysis method based on feature matching fusion |
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 |