CN111445165A - An online classification and early warning evaluation method for tunnel structure health monitoring - Google Patents

Abstract

The invention discloses an online graded early-warning evaluation method for tunnel structure health monitoring, which utilizes high-frequency data of automatic monitoring and realizes real-time evaluation and graded early-warning of tunnel structure safety through an information system, so as to timely know and prevent the safety status of tunnel structures. Provide decision-making advice for sexual maintenance and major and medium repairs. The invention combines the intelligent monitoring technology of the health of the tunnel structure, realizes the online graded early warning based on multi-sensor, multi-index and multi-level data fusion, and obviously reduces the possibility of false alarms or missed alarms; the method of the invention is developed into a set of convenient The software system with good operation and interaction can realize online real-time evaluation and early warning, improve the analysis and response efficiency and intelligence of the tunnel health monitoring system, reduce the manual workload, and reduce the cost for the later operation of the tunnel. Therefore, the present invention is an evaluation method with high feasibility, simple programming and real-time updating, which can ensure the healthy operation of the tunnel in the later stage.

Description

An online classification and early warning evaluation method for tunnel structure health monitoring

technical field

The invention belongs to the technical field of tunnel structure safety monitoring engineering, and in particular relates to an online graded early warning evaluation method for tunnel structure health monitoring.

Background technique

Tunnel structure health monitoring refers to the long-term online monitoring of various mechanical responses of the tunnel structure during the operation period by installing sensors at key positions of the tunnel, so as to realize real-time evaluation and early warning of the health status of the tunnel structure, and to ensure the safe operation and maintenance of the tunnel. Provide guidance. How to realize real-time evaluation and intelligent automatic warning of tunnel structural health status with the help of a large amount of monitoring data collected by the structural health monitoring system is a key problem that needs to be solved urgently. At present, relevant research is still lacking, and there is no corresponding standard at home and abroad. .

Existing tunnel health monitoring systems often use a single sensor or a single evaluation index for evaluation and early warning. Due to factors such as the sensor's own stability and external environmental interference, false alarms or missed alarms will inevitably occur. In view of the great limitations of using a single indicator for early warning, in order to effectively solve this problem, it is necessary to realize real-time comprehensive evaluation and early warning based on multi-indicator data fusion.

Real-time comprehensive evaluation and early warning of tunnel structure health status based on multi-index data fusion. The existing research mainly includes efficacy coefficient method "Li Ming, Chen Weizhong, Yang Jianping. Early warning research on tunnel structure health monitoring system based on efficacy coefficient method [ J].Geotechnical Mechanics, 2015,36(S2):729-736, Distance Discriminant Method "Li Changjun, Chen Weizhong, Li Ming, etc. Application of Weighted Distance Discriminant Analysis Method in Tunnel Health Monitoring System Early Warning[J]. Highway, 2019, 39(06):162-168” et al. Both the efficacy coefficient method and the distance discrimination method force the data of multiple sensors with various monitoring indicators to be assigned to the efficacy coefficient or forcibly normalized, which can realize the data fusion of any number of sensors in form, but the shortcomings are very obvious. , it cannot classify the sensors of various monitoring indicators, forcibly confuse various structural responses, cannot reflect the main factors affecting the health of the tunnel structure, and cannot evaluate the tunnel structure in depth from different dimensions such as internal force, deformation, and durability. Health status.

In view of this, it is necessary to propose a new real-time evaluation and early warning method: it can not only overcome the problem of false alarms or missed alarms in early warning of a single sensor or a single evaluation index, but also overcome the existing online methods such as the efficacy coefficient method and the distance discrimination method. The defects of real-time comprehensive evaluation and data forced integration with early warning methods, no distinction between primary and secondary, and inability to evaluate in depth.

SUMMARY OF THE INVENTION

In view of the above problems, the present invention provides an online graded early-warning evaluation method for tunnel structure health monitoring, which utilizes the high-frequency data of automatic monitoring and realizes "single sensor→single evaluation index→multiple evaluation indexes→...→ The multi-index, multi-level and multi-dimensional real-time evaluation and graded early warning of the entire section/entire section, and the selection of reasonable weights and membership functions can better reflect the impact of various factors on the health of the tunnel structure. It truly realizes the scientific and organic fusion analysis of a large amount of monitoring data of multiple indicators and multiple sensors, and provides scientific and reasonable decision-making suggestions for timely knowledge of the safety status of the tunnel structure, preventive maintenance and major and medium repairs.

An online graded early-warning evaluation method for tunnel structure health monitoring, comprising the following steps:

(1) Determine the health evaluation index system of shield tunnel structure, and divide the system into multiple levels according to the actual situation;

(2) Determine the early warning interval and early warning level of the bottom evaluation index of the system;

(3) Establish the early warning level fuzzy relation equation of each level of evaluation index;

(4) Determine the weight of each evaluation index at each level;

(5) According to the above-mentioned fuzzy relationship equation of early warning level and the weight of each evaluation index, calculate the membership degree of each evaluation index for each early warning level from the bottom layer to the top, until the calculation of the health status of shield tunnel structure for each early warning is obtained. The degree of membership of the grade, the warning level with the highest degree of membership is taken as the decision-making suggestion for the safety evaluation of the tunnel structure and the online real-time warning.

Further, in the step (1), after determining the evaluation index system of the structural health of the shield tunnel, the evaluation index system is divided into a first level, that is, the structural health of the shield tunnel is divided into several evaluation indexes to represent; and then The evaluation index system is then divided into a second level, that is, the evaluation index of the first level is further divided into several sub-categories of evaluation indexes, and so on, until the evaluation index system is divided into multiple levels.

Further, the early warning levels in the step (2) are divided into four levels: green, blue, orange, and red, and the four levels correspond to four groups of early warning intervals respectively, and the four groups of early warning intervals are combined with actual engineering according to existing engineering specifications. The status is determined, in which the green interval represents that the corresponding evaluation index is in the safe range; the blue interval represents that the corresponding evaluation index is too large and needs attention; the orange interval represents that the corresponding evaluation index is too large and cannot be ignored and needs to be paid close attention; the red interval represents the corresponding evaluation The indicators have reached the limit, and relevant experts must be organized to conduct on-site investigation and processing.

Further, for any evaluation index X of the current level in the step (3), if the evaluation index X is subdivided into n evaluation indexes, then the early warning level fuzzy relation equation expression of the evaluation index X of this level is as follows:

Among them: a _i indicates that the evaluation index X is subdivided into the weight of the i-th evaluation index among the n evaluation indicators, and l _i1 to l _i4 are the membership of the i-th evaluation index to the four levels of green, blue, orange, and red, respectively. degree, b ₁ to b ₄ are the degrees of membership of the evaluation index X to the four grades of green, blue, orange, and red, respectively, i is a natural number and 1≤i≤n.

Further, in the step (4), the three-scale method or the nine-scale method is used to determine the weight of each evaluation index at each level.

Further, in the step (5), the membership degree of each evaluation index for each early warning level is calculated layer by layer from the bottom layer, and the monitoring value of each evaluation index at the bottom layer needs to be collected in real time through sensors; For any evaluation index, if the current monitoring value of the evaluation index is x, its membership degree to each early warning level is determined by the following criteria:

If the warning interval corresponding to the warning level is [0, c ₁ ], the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ , c ₂ ] and c ₁ c ₂ <0, the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ , c ₂ ] and c ₁ c ₂ > 0, the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ ,c ₂ ]&[c ₃ ,c ₄ ] and c ₃ <c ₄ <0<c ₁ <c ₂ , then the lowest-level evaluation index belongs to the warning level The degree f(x) is:

If the warning interval corresponding to the warning level is [c ₁ ,+∞), the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is (-∞,c ₁ ]&[c ₂ ,+∞) and c ₁ c ₂ <0, then the membership degree f(x) of the lowest evaluation index to the warning level is:

Among them: c ₁ , c ₂ , c ₃ , and c ₄ are the thresholds of the warning interval, and F() is a Boolean function, that is, when the relational expression in ( ) is satisfied, the function value is 1, otherwise the function value is 0.

Further, when the monitoring value of a bottom evaluation index is collected by multiple sensors, the degree of membership f(x) of the evaluation index to each early warning level is:

Among them: m is the number of sensors, x ₁ , x ₂ ,..., x _m are the monitoring values collected by the m sensors, respectively.

Combined with the intelligent monitoring technology of tunnel structure health, the invention realizes on-line graded early warning based on multi-sensor, multi-index and multi-level data fusion, and obviously reduces the possibility of false alarms or missed alarms. The method of the invention is developed into a set of software system that is easy to operate and has good interactivity, which can realize online real-time evaluation and early warning, improve the analysis and response efficiency and intelligent degree of the tunnel health monitoring system, reduce the manual workload, and reduce the operation of the tunnel in the later stage. cost. Therefore, the present invention is an evaluation method with high feasibility, simple programming, small calculation amount and real-time updating, which can ensure the healthy operation of the tunnel in the later stage.

Description of drawings

FIG. 1 is a schematic diagram of a hierarchical model tree of the tunnel structure health condition evaluation index system of the present invention.

Figure 2(a) is a schematic diagram of the membership function corresponding to the warning interval [0, x ₁ ].

Figure 2(b) is a schematic diagram of the membership function corresponding to the warning interval [x ₁ , x ₂ ] (x ₁ x ₂ <0).

Figure 2(c) is a schematic diagram of the membership function corresponding to the warning interval [x ₁ , x ₂ ] (x ₁ x ₂ >0).

Figure 2(d) is a schematic diagram of the membership function corresponding to the warning interval [x ₁ , x ₂ ]&[x ₃ , x ₄ ].

Figure 2(e) is a schematic diagram of the membership function corresponding to the warning interval [x ₁ , +∞).

Figure 2(f) is a schematic diagram of the membership function corresponding to the warning interval (-∞, x ₁ ]&[x ₂ , +∞).

Figure 3(a) is a schematic diagram of the layout of the embedded sensor in section A.

Figure 3(b) is a schematic diagram of the layout of the surface sensor in section A.

Figure 3(c) is a schematic diagram of the layout of the monitoring points of the static level in the monitoring section of the tunnel.

Figure 4(a) is a schematic diagram of the monitoring data of the seam opening.

Figure 4(b) is a schematic diagram of the vault deformation monitoring data.

Figure 4(c) is a schematic diagram of monitoring data of segment tilt deflection.

Figure 4(d) is a schematic diagram of the monitoring data of longitudinal relative uneven settlement.

Figure 4(e) is a schematic diagram of the concrete strain monitoring data.

Figure 4(f) is a schematic diagram of the monitoring data of the internal force of the steel bar.

Figure 4(g) is a schematic diagram of the strain monitoring data on the bottom surface of the lane slab.

FIG. 5 is a schematic diagram of a hierarchical model tree of a tunnel structure health evaluation index system according to an embodiment of the present invention.

FIG. 6 is a schematic diagram of a graded early warning result at a certain moment in an embodiment of the present invention.

Figure 7 is a schematic diagram of the real-time evaluation results of the overall safety of the structure of section A within a period of time.

Detailed ways

In order to describe the present invention more specifically, the technical solutions of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The online grading early warning evaluation method for tunnel structure health monitoring of the present invention includes the following steps:

Step 1: Determine the index system for evaluating the structural health of the shield tunnel, and classify and divide each index into multiple levels according to the actual situation.

As shown in Figure 1, to determine the evaluation index system of shield tunnel structure health status, first divide the shield tunnel structure health status evaluation index system into the first level, and divide it into a, b, ..., n indicators, Then, the second level is divided into the evaluation index system of shield tunnel structure health status, and all indexes of the first level are divided into a ₁ , a ₂ , ..., a _r , b ₁ , b ₂ , ..., b _s , ..., n ₁ , n ₂ , ..., _nt , and so on, until the index system for evaluating the structural health of the shield tunnel is divided into S levels.

Step 2: Determine the warning interval and warning level of the bottom evaluation index.

2.1 The warning level is divided into four intervals: green, blue, orange and red. It is stipulated that when the internal force deformation of the structure does not exceed the blue warning value, it is the green interval, and the blue warning value exceeds the orange warning value. The blue warning value is blue, and the orange warning value exceeds the orange warning value. The value does not exceed the red warning value for the orange interval, and the red warning value exceeds the red interval. The green interval represents that the internal force deformation of the structure is in the safe range; the blue interval represents that the internal force deformation of the structure is large and needs attention; the orange interval represents that the internal force deformation of the structure cannot be ignored and needs to be paid close attention; the red interval represents that the internal force deformation of the structure has reached the bearing capacity of the structure If the force limit or the structural service limit is exceeded, relevant experts must be organized to conduct on-site investigation and processing.

2.2 On the basis of referring to the existing relevant norms or related research, first determine the specific value of the blue early warning value or the red early warning value, and then specify the proportional relationship between the early warning values at all levels according to the situation, and the blue early warning value can be specified. a% of the red early warning value, and the orange early warning value is b% of the red early warning value (where 0<a<b<100).

Step 3: Establish a fuzzy relational equation system to determine the early warning level of each level of evaluation index.

The membership degrees of all lower-level indicators to each early warning level are passed through the fuzzy relation equation:

The calculation is transformed into the membership degree of the index of the previous layer to each early warning level, and the level is upward, and finally the membership degree of the overall structural health of the tunnel to each early warning level is obtained. where a _i is the normalized weight of the corresponding factor, l _ij and b _ij are the degree of membership of the i-th factor to the j-th early warning level, * is the fuzzy operator (the present invention adopts the most commonly used weighted average operator), l _1j and b ₁ correspond to the green early warning level, l _2j and b ₂ correspond to the blue early warning level, l _3j and b ₃ correspond to the orange early warning level, and l _4j and b ₄ correspond to the red early warning level.

Step 4: Use the three-scale method to determine the weight of each index at each level.

4.1 Determining the relative importance of indicators to decision-making objectives in the three-scale method can be expressed as:

4.2 According to specific problems, a pairwise comparison matrix C can be generated:

4.3 The optimal transfer matrix O calculated based on the pairwise comparison matrix C:

in:

4.4 Convert the optimal transfer matrix into a consistency judgment matrix A:

故c_ij＝c_imc_mj，依据一致性矩阵的定义矩阵A无需进行一致性检验。where: a _ij =exp(o _ij ), since

Therefore, c _ij =c _im c _mj , and the matrix A does not need to be checked for consistency according to the definition of the consistency matrix.

After the consistency judgment matrix is obtained, its eigenvalues and corresponding eigenvectors can be obtained, and the eigenvector corresponding to the largest eigenvalue can be normalized to obtain the weight of each index to the decision target.

Step 5: Construct a membership function for each evaluation index at the bottom layer, perform weighted average fusion of multiple monitoring data in a section of each evaluation index at the bottom layer, and substitute the weighted average fusion value into the membership function to calculate the corresponding evaluation index for each Membership of the warning level.

The membership function of the present invention is constructed as a normal distribution type, and obtains:

Among them: f(x) is the membership degree to the specified warning level when the structural response value of a certain evaluation index is x, when the monitoring value x of a single sensor is known, then f(x) is the evaluation at the measuring point of the sensor The membership degree of the indicator to a certain warning level; F(r) is a Boolean function, if r is established, then F(r)=1; otherwise, F(r)=0. When determining the degree of membership, you should first look at the form of the warning interval, and then determine the corresponding function calculation formula. The function image form of each warning interval is shown in Figure 2(a) to Figure 2(f).

The formula for calculating the weighted average fusion value of multiple sensors is as follows:

Step 6: Embed the above algorithm into the automatic monitoring software to realize online multi-level and hierarchical early warning, and timely discover hidden dangers of tunnel structure safety.

In the following embodiments, the sensor arrangement of a tunnel section A is shown in Figures 3(a) to 3(c), of which Figure 3(a) is the installation diagram of the embedded sensor; Figure 3(b) is the installation of the surface sensor Fig. 3(c) is the sensor layout of the longitudinal uneven settlement monitoring section spanning section A, including 9 fiber grating static levels, and a measuring point is arranged every 20 meters. The sensor monitoring data of all sensors in section A in a certain period are shown in Figure 4(a) to Figure 4(g).

Based on the sensor monitoring data of section A, the online graded early warning evaluation of the structural health status of the section is carried out.

Step a: Define the health status of the tunnel structure into two levels. The first level is structural deformation and structural internal force. In the second level, the structural deformation is subdivided into joint opening, longitudinal uneven settlement, vault convergence and segment inclination. Deflection, the internal force of the structure is subdivided into steel internal force, concrete strain and lane slab strain to form a hierarchical model tree of the evaluation index system as shown in Figure 5.

Step b: With reference to relevant specifications and relevant scientific research achievements, determine the early warning level of each specific indicator and the division result of the early warning interval in this embodiment, as shown in Table 1.

Table 1

Step c: According to the hierarchical model tree in Figure 5, establish a fuzzy relational equation system for calculating the early warning levels of the evaluation indicators at each level:

Step d: Use the three-scale method to determine the weight of each index in the hierarchical model tree.

For the first level, the structural deformation is more important than the internal force of the structure, and the three-scale method can be used to obtain the respective weights of 0.73 and 0.27, respectively. For the second level, the respective weights of the four indicators under the structural deformation are 0.45, 0.28, 0.17, and 0.1; the respective weights of the three indicators under the structural internal force are 0.56, 0.29, and 0.15.

Step e: Calculate the membership degree of each level of evaluation index to each early warning level according to the membership degree function form corresponding to FIG. 2(a)-FIG. 2(f). First, perform weighted average fusion of multiple monitoring data in a section of each evaluation index at the bottom layer, and substitute the weighted average fusion value into the membership function to calculate the membership degree of the corresponding evaluation index to each early warning level; The membership degree is substituted into the fuzzy relation equation system, and the calculation is carried out layer by layer. Finally, the membership degree of all evaluation indicators for each early warning level can be obtained. Figure 6 shows the hierarchical early warning of this example at 12:00 noon on September 16. result.

Step f: The above algorithm is implanted into the automatic monitoring software, and the continuous real-time evaluation result of the overall safety of the section A structure within a period of time is calculated, as shown in Figure 7.

The above description of the embodiments is for the convenience of those of ordinary skill in the art to understand and apply the present invention. It will be apparent to those skilled in the art that various modifications to the above-described embodiments can be readily made, and the general principles described herein can be applied to other embodiments without inventive effort. Therefore, the present invention is not limited to the above-mentioned embodiments, and improvements and modifications made by those skilled in the art according to the disclosure of the present invention should all fall within the protection scope of the present invention.

Claims (8)

1. An online grading early warning evaluation method for tunnel structure health monitoring, comprising the following steps: (1) Determine the health evaluation index system of shield tunnel structure, and divide the system into multiple levels according to the actual situation; (2) Determine the early warning interval and early warning level of the bottom evaluation index of the system; (3) Establish the early warning level fuzzy relation equation of each level of evaluation index; (4) Determine the weight of each evaluation index at each level; (5) According to the above-mentioned fuzzy relationship equation of early warning level and the weight of each evaluation index, calculate the membership degree of each evaluation index for each early warning level from the bottom layer to the top, until the calculation of the health status of shield tunnel structure for each early warning is obtained. The degree of membership of the grade, the warning level with the highest degree of membership is taken as the decision-making suggestion for the safety evaluation of the tunnel structure and the online real-time warning. 2. The online grading and early warning evaluation method for tunnel structure health monitoring according to claim 1, characterized in that: in the step (1), after determining the shield tunnel structure health condition evaluation index system, the first evaluation index system is performed. Hierarchical division, that is, the health status of shield tunnel structure is divided into several evaluation indexes to represent; and then the evaluation index system is divided into a second level, that is, the first level of evaluation indexes is further divided into several sub-category evaluation indexes, And so on, until the evaluation index system is divided into multiple levels. 3. The tunnel structure health monitoring online grading early warning evaluation method according to claim 1, is characterized in that: the early warning level in the described step (2) is divided into four levels of green, blue, orange and red, and the four levels are respectively Corresponding to four groups of early-warning intervals, the four groups of early-warning intervals are determined according to the existing engineering specifications and actual engineering conditions. The green interval represents that the corresponding evaluation index is in the safe range; the blue interval represents that the corresponding evaluation index is too large and needs attention; the orange interval It means that the corresponding evaluation index is too large and cannot be ignored, and needs to be paid close attention; the red interval means that the corresponding evaluation index has reached the limit, and relevant experts must be organized to conduct on-site investigation and processing. 4. The online grading and early warning evaluation method for tunnel structure health monitoring according to claim 1, characterized in that: in the step (3), for any evaluation index X of the current level, if the evaluation index X is subdivided into n For each evaluation index, the early warning level fuzzy relation equation of the evaluation index X at this level is expressed as follows:

Among them: a _i indicates that the evaluation index X is subdivided into the weight of the i-th evaluation index among the n evaluation indicators, and l _i1 to l _i4 are the membership of the i-th evaluation index to the four levels of green, blue, orange, and red, respectively. degree, b ₁ to b ₄ are the degrees of membership of the evaluation index X to the four grades of green, blue, orange, and red, respectively, i is a natural number and 1≤i≤n. 5 . The online grading and early warning evaluation method for tunnel structure health monitoring according to claim 1 , wherein: in the step (4), a three-scale method or a nine-scale method is used to determine the weight of each evaluation index at each level. 6 . 6. The online grading early warning evaluation method for tunnel structure health monitoring according to claim 1, characterized in that: in the step (5), the degree of membership of each evaluation index for each early warning level is calculated from the bottom layer up layer by layer. , the monitoring value of each evaluation index at the bottom must be collected in real time through the sensor; for any evaluation index at the bottom, if the current monitoring value of the evaluation index is x, the following criteria are used to determine the degree of membership of each early warning level: If the warning interval corresponding to the warning level is [0, c ₁ ], the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ , c ₂ ] and c ₁ c ₂ <0, the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ , c ₂ ] and c ₁ c ₂ > 0, the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is [c ₁ ,c ₂ ]&[c ₃ ,c ₄ ] and c ₃ <c ₄ <0<c ₁ <c ₂ , then the lowest-level evaluation index belongs to the warning level The degree f(x) is:

If the warning interval corresponding to the warning level is [c ₁ ,+∞), the membership degree f(x) of the lowest evaluation index to the warning level is:

If the warning interval corresponding to the warning level is (-∞,c ₁ ]&[c ₂ ,+∞) and c ₁ c ₂ <0, then the membership degree f(x) of the lowest evaluation index to the warning level is:

Among them: c ₁ , c ₂ , c ₃ , and c ₄ are the thresholds of the warning interval, and F() is a Boolean function, that is, when the relational expression in ( ) is satisfied, the function value is 1, otherwise the function value is 0. 7. The online grading early-warning evaluation method for tunnel structure health monitoring according to claim 6, characterized in that: when a certain bottom-level evaluation index collects monitoring values by multiple sensors, the degree of membership of the evaluation index to each early-warning level f(x) is:

Among them: m is the number of sensors, x ₁ , x ₂ ,..., x _m are the monitoring values collected by the m sensors, respectively. 8. The online grading and early warning evaluation method for tunnel structure health monitoring according to claim 1, characterized in that: the method combines the intelligent monitoring technology of tunnel structure health, and realizes online multi-sensor, multi-index, and multi-level data fusion based on Graded early warning significantly reduces the possibility of false alarms or missed alarms.

Images