Deformation dynamic prediction and early warning method for tunnel under construction based on multi-source data fusion

Abstract

The present invention relates to a deformation dynamic prediction and early warning method for a tunnel under construction based on multi-source data fusion, and mainly solves the technical problem of large difficulty in evaluating tunnel safety states using macroscopic analysis. The present invention comprises: step 1, dividing dynamic prediction and early warning of an entire tunnel construction process into three stages; step 2, determining a safety level standard framework, and constructing an index system; step 3, constructing identification intervals of each index of an index layer in different safety states; step 4, specifying a relative weight distribution of each index; step 5, calculating distances from an interval to the identification intervals for different safety states on the basis of an interval distance formula, and generating interval distance vectors; step 6, converting the interval distance vectors into similarity vectors on the basis of a similarity formula, carrying out normalization processing, and constructing an analysis matrix; step 7, implementing same-level index data fusion by calculating a product of a relative weight vector of each index of the same layer and the analysis matrix layer by layer, and obtaining a safety level probability distribution of an entire section of a tunnel at a time point; and step 8, dynamically tracking a dominant state of a tunnel safety level, and performing dynamic prediction and early warning on the entire tunnel construction process.

Description

A dynamic prediction and early warning method for construction tunnel deformation based on multi-source data fusion Technical Field

[0001] The present invention relates to the field of tunnel construction early warning, and in particular to a construction tunnel deformation dynamic prediction and early warning method based on multi-source data fusion. Background Art

[0002] Highway tunnel projects are characterized by complex geological conditions and uncertain factors. It is difficult to evaluate the safety status of tunnels using a macro-analysis approach. The accuracy often deviates from the actual situation due to incomplete considerations, and a large amount of information needs to be collected, which is not universal. For example, CN202310828365.2 involves an intelligent decision-making method and auxiliary platform for early warning and prevention of geological disasters in tunnel construction. The decision-making method includes collecting tunnel data, identifying unfavorable geology, three-dimensional geological modeling, disaster monitoring and early warning, and disaster prevention and control decisions. Therefore, it is necessary to use a quantitative analysis method to calculate the probability distribution of tunnel safety levels and to perform real-time dynamic prediction and early warning based on the probability distribution time series evolution model. The surrounding rock level and monitoring and measurement measured data are used as media to implement a graded, dynamic, and comprehensive analysis of the entire tunnel construction process, thereby predicting the development trend of the tunnel safety status and providing a real and effective basis for the dynamic design and construction of the tunnel.

[0003] Summary of the Invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned existing technologies and to provide a dynamic prediction and early warning method for construction tunnel deformation based on multi-source data fusion. The surrounding rock level and monitoring measurement data are used as media to implement graded, dynamic and comprehensive analysis of the entire process before the secondary lining construction of the tunnel, and prediction and early warning are carried out based on the safety probability distribution time series evolution model.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion includes the following steps:

[0007] Step 1: Divide the dynamic prediction and early warning of the entire tunnel construction process into three stages:

[0008] 1) Prediction in the unexcavated stage: At this time, the monitored section has not been excavated. The advanced geological prediction method is used to test the groundwater development degree and surrounding rock integrity index of the monitored section, and the surrounding rock stability is comprehensively predicted in combination with the exploration surrounding rock strength.

[0009] 2) Prediction during the excavation and support phase: First, when excavating to the monitoring section, a stability assessment is conducted based on the groundwater conditions at the tunnel face, the integrity of the surrounding rock mass, the uniaxial saturated compressive strength, and monitoring measurements. Second, after the monitoring section is supported, a dynamic stability assessment is conducted based on the surrounding rock mass grade and monitoring measurements, leading to prediction and early warning.

[0010] 3) Termination stage: The prediction and early warning activities are terminated after the secondary lining construction of the monitoring section is completed.

[0011] Step 2: Determine the security level framework standard and build an indicator system;

[0012] Step 3: Construct the identification intervals of each indicator in the indicator layer in different security states;

[0013] Step 4: Specify the relative weight distribution of each indicator;

[0014] Step 5: Convert the field measured data into interval numbers, and calculate the distance from the interval to the identification interval of different safety states according to the interval distance formula to generate the interval distance vector;

[0015] Step 6: Convert the interval distance vector into a similarity vector according to the interval similarity formula, normalize it, and construct an analysis matrix;

[0016] Step 7: By calculating the relative weight vectors of each indicator at the same level and multiplying them by the analysis matrix, the data of indicators at the same level are integrated to obtain the probability distribution of the safety level of the entire tunnel section at a certain point in time. The safety level at each time point is represented by a probability distribution of five levels: {Level 1, Level 2, Level 3, Level 4, Level 5} = (q1, q2, q3, q4, q5). Among them, q1 + q2 + q3 + q4 + q5 = 1. The level with the highest probability is called the explicit state, which is the actual safety level of the tunnel. The other four safety levels are called implicit states.

[0017] Step 8: Analyze the temporal trends in the probability distribution of tunnel safety levels at different time points, dynamically track the dominant state of the tunnel safety level, and implement dynamic prediction and early warning throughout the tunnel construction process. Tunnel instability and failure manifests as a dynamic progression from safety level 1 to safety level 5, specifically as the probability of the previous level decreases and the probability of the next level increases, alternating between them. That is, over time, the tunnel safety level changes from dominant level 1 to dominant level 2, then to dominant level 3, then to dominant level 4, then to dominant level 5. The first warning is issued when the tunnel reaches level 3, the second warning when it reaches level 4, and the immediate suspension of construction is required when it reaches level 5.

[0018] Furthermore, the safety level standard framework is divided into {Level 1, Level 2, Level 3, Level 4, Level 5}, where Level 1 is the safest state and Level 5 is the most dangerous state.

[0019] Furthermore, the indicator system includes five levels, the first level being the tunnel section safety level.

[0020] The secondary indicators are set as follows:

[0021] Tunnel section safety level = {surrounding rock level, monitoring measurement};

[0022] The three-level indicators are set as follows:

[0023] Surrounding rock grade = {groundwater conditions, uniaxial saturated compressive strength, integrity index};

[0024] Monitoring measurement = {displacement, velocity}

[0025] The four-level indicators are set as follows:

[0026] Displacement = {vault subsidence, peripheral displacement};

[0027] Rate = {vault subsidence rate, peripheral displacement rate};

[0028] The five-level indicators are set as follows:

[0029] The vault settlement, peripheral displacement, vault settlement rate, and peripheral displacement rate are respectively used as control indicators at each measuring point.

[0030] Furthermore, the five levels are specifically configured as follows:

[0031] Groundwater status identification interval division: poor, relatively developed, developed, abundant, extremely abundant = [0,1], [1,2], [2,3], [3,4], [4,5];

[0032] Uniaxial saturated compressive strength identification interval division: very soft rock, soft rock, relatively soft rock, hard rock, very hard rock = [0,5], [5,15], [15,30], [30,60], [60,120];

[0033] Integrity index identification interval division: extremely broken, broken, relatively broken, relatively complete, complete = [0, 0.15], [0.15, 0.35], [0.35, 0.55], [0.55, 0.75], [0.75, 1];

[0034] The identification intervals of vault subsidence and peripheral displacement are divided into [0,1 / 5U0], [1 / 5U0,2 / 5U0], [2 / 5U0,3 / 5U0], [3 / 5U0,4 / 5U0], and [4 / 5U0,U0], where U0 is the design reserved deformation.

[0035] The recognition intervals for the vault subsidence and peripheral displacement rates are divided into [0, 1 / 5v0], [1 / 5v0, 2 / 5v0], [2 / 5v0, 3 / 5v0], [3 / 5v0, 4 / 5v0], and [4 / 5v0, v0], where v0 is the limit rate. For example, when v0 is set to 1 mm / d, the rate recognition intervals are set as follows: [0, 0.2], [0.2, 0.4], [0.4, 0.6], [0.6, 0.8], and [0.8, 1];

[0036] Furthermore, the relative weights of the indicators are allocated as follows:

[0037] Table 2 Configuration table of weights and support coefficients of indicators at all levels

[0038] The weight distribution of the five-level indicators is evenly distributed according to the number of measuring points.

[0039] Furthermore, the interval distance calculation formula is as follows:

[0040] Let A = [a1, a2] and B = [b1, b2] be two interval numbers, then the square of their distance D 2 for:

[0041] Interval distance The real number a is converted into an interval number represented as [a, a]. The distance from the original data to each recognition interval is calculated according to the above formula, and the interval distance vector D = (d1, d2, d3, d4, d5) is obtained.

[0042] Furthermore, the interval similarity formula calculation method is as follows.

[0043] Assume A = [a1, a2], B = [b1, b2] are two interval numbers, then the similarity S(A, B) of the interval numbers A and B is:

[0044] Where α>0 is the support coefficient, and D(A, B) is the distance between intervals A and B. The support coefficient α is mainly used to adjust the dispersion of the generated similarity values. The interval distance vector is converted to a similarity vector S = (s1, s2, s3, s4, s5) using the interval similarity formula.

[0045] Furthermore, the analysis matrix is ​​composed of a plurality of normalized similarity vectors of the same-level indicators.

[0046] Assume that a certain level contains m indicators, and the weight vector of these m indicators is w=(w1,w2,...,w m ), the normalized similarity vector of the jth index is r j =(a j1, a j2 , a j3 , a j4 , a j5 ), j=1,2,...,m. The matrix Q composed of the calculation results of the same-level indicators i for:

[0047] Furthermore, the data fusion is calculated using the following formula:

[0048] The calculation result Q of the previous level i+1 For: Q i+1 =W*Q i

[0049] After layer-by-layer fusion, the final security state vector (q1, q2, q3, q4, q5) for a particular tunnel section is obtained. Q1, q2, q3, q4, and q5 correspond to security levels 1, 2, 3, 4, and 5, respectively. The security level at each time point is represented by a five-level probability distribution: {Level 1, Level 2, Level 3, Level 4, Level 5} = (q1, q2, q3, q4, q5). Q1 + q2 + q3 + q4 + q5 = 1. The level with the highest probability is called the explicit state, i.e., the actual security level of the tunnel. The other four security levels are called implicit states.

[0050] Furthermore, the probability distribution of tunnel safety levels at different time points was analyzed over time, and the dominant state of the tunnel safety level was dynamically tracked, enabling dynamic prediction and early warning throughout the entire tunnel construction process. Tunnel instability and failure manifests itself as a dynamic progression from safety level 1 to safety level 5, specifically as the probability of each level decreases and the probability of each level increases, alternating between levels. Over time, the tunnel safety level changes from dominant level 1 to dominant level 2, then to dominant level 3, then to dominant level 4, finally to dominant level 5. A first warning is issued when the tunnel reaches level 3, a second warning at level 4, and an immediate construction halt at level 5.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1) Dynamic tracking, prediction and early warning of the entire process from the unexcavated stage of the tunnel face to the excavation stage and then to the secondary lining closure are realized.

[0053] 2) Calculate the probability distribution of tunnel safety levels based on measured tunnel geological data and dynamic monitoring data. Comprehensively determine the safety level by integrating geological factors and monitoring data, effectively avoiding the influence of subjective factors. This approach also features a simple framework, easy data acquisition, and greater universality.

[0054] 3) Based on the probability distribution time series evolution model, it is possible to infer future development trends and make predictions and early warnings. This method can reflect the dynamic changes in the tunnel safety status through continuous probability distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] FIG1 is a flow chart of an implementation of the present invention;

[0056] FIG2 is a flow chart of the dynamic prediction timing of the present invention;

[0057] FIG3 is a diagram of an indicator framework system of the present invention;

[0058] FIG4 is a probability distribution time series evolution early warning model of the present invention. DETAILED DESCRIPTION

[0059] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0060] This embodiment provides a method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion, as shown in Figure 1. The method includes the following steps:

[0061] S1. As shown in Figure 2, the dynamic early warning of the entire tunnel construction process is divided into three stages:

[0062] 1) Prediction in the unexcavated stage: At this time, the monitored section has not been excavated. The advanced geological prediction method is used to test the groundwater development degree and surrounding rock integrity index of the monitored section, and the surrounding rock stability is comprehensively predicted in combination with the exploration surrounding rock strength.

[0063] 2) Prediction during the excavation and support phase: First, when excavating to the monitoring section, a stability assessment is conducted based on the groundwater conditions at the tunnel face, the integrity of the surrounding rock mass, the uniaxial saturated compressive strength, and monitoring measurements. Second, after the monitoring section is supported, a dynamic stability assessment is conducted based on the surrounding rock mass grade and monitoring measurements, leading to prediction and early warning.

[0064] 3) Termination stage: The prediction and early warning activities are terminated after the secondary lining construction of the monitoring section is completed.

[0065] S2. Determine the security level framework and construct the indicator system as shown in Figure 3;

[0066] The safety level standard framework is divided into {Level 1, Level 2, Level 3, Level 4, Level 5}, where Level 1 is the safest state and Level 5 is the most dangerous state.

[0067] The index system consists of five levels. The first level is the tunnel section safety level, and the second level indicators include surrounding rock level and monitoring measurement.

[0068] The secondary indicators are set as follows:

[0069] Tunnel section safety level = {surrounding rock level, monitoring measurement};

[0070] The three-level indicators are set as follows:

[0071] Surrounding rock grade = {groundwater conditions, uniaxial saturated compressive strength, integrity index};

[0072] Monitoring measurement = {displacement, velocity};

[0073] The four-level indicators are set as follows:

[0074] Displacement = {vault subsidence, peripheral displacement};

[0075] Rate = {vault subsidence rate, peripheral displacement rate};

[0076] The five-level indicators are set as follows:

[0077] The vault settlement, peripheral displacement, vault settlement rate, and peripheral displacement rate are respectively used as control indicators at each measuring point.

[0078] S3. Construct the identification intervals of each indicator in the indicator layer in different security states;

[0079] Furthermore, the five levels are specifically configured as follows:

[0080] Groundwater status identification interval division: poor, relatively developed, developed, abundant, extremely abundant = [0,1], [1,2], [2,3], [3,4], [4,5];

[0081] Uniaxial saturated compressive strength identification interval division: very soft rock, soft rock, relatively soft rock, hard rock, very hard rock = [0,5], [5,15], [15,30], [30,60], [60,120];

[0082] Integrity index identification interval division: extremely broken, broken, relatively broken, relatively complete, complete = [0, 0.15], [0.15, 0.35], [0.35, 0.55], [0.55, 0.75], [0.75, 1];

[0083] The identification intervals of vault subsidence and peripheral displacement are divided into [0,1 / 5U0], [1 / 5U0,2 / 5U0], [2 / 5U0,3 / 5U0], [3 / 5U0,4 / 5U0], and [4 / 5U0,U0], where U0 is the design reserved deformation.

[0084] The recognition intervals for the vault subsidence and peripheral displacement rates are divided into [0, 1 / 5v0], [1 / 5v0, 2 / 5v0], [2 / 5v0, 3 / 5v0], [3 / 5v0, 4 / 5v0], and [4 / 5v0, v0], where v0 is the limit rate. For example, when v0 is set to 1 mm / d, the rate recognition intervals are set as follows: [0, 0.2], [0.2, 0.4], [0.4, 0.6], [0.6, 0.8], and [0.8, 1];

[0085] S4, relative weight distribution of each indicator;

[0086] Table 3. Weights and support coefficients of indicators at all levels

[0087] The weight distribution of the five-level indicators is evenly distributed according to the number of measuring points.

[0088] S5. Convert the field measured data into interval numbers, and calculate the distance between the interval and the identification intervals of different safety states according to the interval distance formula to generate an interval distance vector;

[0089] The interval distance calculation formula is as follows:

[0090] Let A = [a1, a2] and B = [b1, b2] be two interval numbers, then the square of their distance D 2 for:

[0091] Interval distance The real number a is converted into an interval number represented as [a, a]. The distance from the original data to each recognition interval is calculated according to the above formula, and the interval distance vector D = (d1, d2, d3, d4, d5) is obtained.

[0092] S6. Convert the interval distance vector into a similarity vector according to the interval similarity formula, normalize it, and construct an analysis matrix;

[0093] The calculation method of interval similarity formula is as follows.

[0094] Assume A = [a1, a2], B = [b1, b2] are two interval numbers, then the similarity S(A, B) of the interval numbers A and B is:

[0095] Where α>0 is the support coefficient, and D(A, B) is the distance between the interval numbers A and B. The support coefficient α is mainly used to adjust the discreteness of the generated similarity values. The interval distance vector is converted to the similarity vector S=(s1, s2, s3, s4, s5) through the interval similarity formula

[0096] Finally, the analysis matrix consists of the normalized similarity vectors of multiple indicators at the same level, as shown below:

[0097] Assume that a certain level contains m indicators, and the weight vector of these m indicators is w=(w1,w2,...,w m ), the normalized similarity vector of the jth index is r j =(a j1 , a j2 , a j3 , a j4 , a j5 ), j=1,2,...,m. The matrix Q composed of the calculation results of the same-level indicators i for:

[0098] S7. Data fusion of indicators at the same level is achieved by calculating the product of the relative weight vector of each indicator at the same level and the analysis matrix layer by layer. The safety probability distribution of a certain section of the tunnel is finally obtained by layer-by-layer data fusion.

[0099] Data fusion is calculated using the following formula:

[0100] The calculation result Q of the previous level i+1 For: Q i+1 =W*Q i

[0101] Through layer-by-layer fusion, the final security state vector (q1, q2, q3, q4, q5) for a particular tunnel section is obtained. Q1, Q2, Q3, Q4, and Q5 correspond to security levels 1, 2, 3, 4, and 5, respectively. Therefore, the security level at each time point is represented by a five-level probability distribution: {Level 1, Level 2, Level 3, Level 4, Level 5} = (q1, q2, q3, q4, q5). Q1 + Q2 + Q3 + Q4 + Q5 = 1. The level with the highest probability is called the explicit state, i.e., the actual security level of the tunnel. The other four security levels are called implicit states.

[0102] S8. Analyze the temporal trends in the probability distribution of tunnel safety levels at different time points, dynamically track the dominant state of tunnel safety levels, and implement dynamic prediction and early warning throughout the tunnel construction process. Tunnel instability and failure manifests itself as a dynamic progression from safety level 1 to safety level 5, as shown in Figure 4. Specifically, the probability of each level decreases while the probability of the next level increases, alternating between them. That is, over time, the tunnel safety level changes from dominant level 1 to dominant level 2, then to dominant level 3, then to dominant level 4, then to dominant level 5. The first warning is issued when the tunnel reaches level 3, the second warning when it reaches level 4, and the immediate halt of construction is required when it reaches level 5.

[0103] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion, characterized in that: The following steps are involved: Step 1: Divide the dynamic prediction and early warning of the whole tunnel construction process into three stages: 1) Prediction in the unexcavated stage: At this time, the monitored section is not excavated. The advanced geological prediction method is used to test the groundwater development degree and surrounding rock integrity index of the monitored section, and the surrounding rock stability is comprehensively predicted in combination with the exploration of surrounding rock strength; 2) Prediction during the excavation and support phase: First, when excavating to the monitoring section, the stability of the monitoring section is evaluated based on the groundwater conditions at the face, the integrity of the surrounding rock, the uniaxial saturated compressive strength, and monitoring measurements; second, after the monitoring section is supported, a dynamic stability evaluation is performed based on the surrounding rock grade and monitoring measurements, and then prediction and early warning are implemented; 3) Termination stage: the prediction and early warning activities are terminated after the secondary lining construction of the monitoring section is closed; Step 2: Determine the safety level standard framework and build an indicator system; Step 3: Construct the identification intervals of each indicator in the indicator layer in different security states; Step 4: Specify the relative weight distribution of each indicator; Step 5: Convert the field measured data into interval numbers, and calculate the distance from the interval to the identification interval of different safety states according to the interval distance formula to generate an interval distance vector; Step 6: Convert the interval distance vector into a similarity vector according to the interval similarity formula, normalize it, and construct an analysis matrix; Step 7: The data of the indicators at the same level are integrated by calculating the relative weight vector of each indicator at the same level layer by layer and multiplying it by the analysis matrix, and the probability distribution of the safety level of a certain section of the tunnel at a certain time point is obtained; the safety level at each time point is represented by the probability distribution of 5 levels: {level 1, level 2, level 3, level 4, level 5} = (q1, q2, q3, q4, q5); wherein q1+q2+q3+q4+q5=1, and the level with the maximum probability is called the explicit state, while the other 4 safety levels are in the implicit state, wherein the explicit state is the real safety level of the tunnel; Step 8: Analyze the temporal trend of probability distribution of tunnel safety level at different time points, dynamically track the explicit status of tunnel safety level, and implement dynamic prediction and early warning for the entire tunnel construction process; tunnel instability and damage is manifested as a dynamic development process in which safety level 1 gradually changes to level 5, which is indicated by the decrease in probability of the previous level and the increase in probability of the next level, alternating in this way.

2. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The safety level standard framework of the step 2 is {level 1, level 2, level 3, level 4, level 5}. As time goes by, the tunnel safety level changes as follows: level 1 explicit state → level 2 explicit state → level 3 explicit state → level 4 explicit state → level 5 explicit state; level 1 is the safest state. When the tunnel is at level 3, the first warning is issued. When it is at level 4, the second warning is issued. When it is at level 5, it is the most dangerous state and construction is required to be stopped immediately.

3. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The indicator system of step 2 includes five levels, the first level is the tunnel section safety level; The second level indicator settings are as follows: Tunnel section safety level = {surrounding rock level, monitoring measurement}; The third level indicator settings are as follows: Surrounding rock grade = {groundwater condition, uniaxial saturated compressive strength, integrity index}; Monitoring measurement = {displacement, velocity}; The fourth level indicators are set as follows: Displacement = {vault subsidence, peripheral displacement}; Rate = {vault sinking rate, peripheral displacement rate}; The fifth level indicators are set as follows: The vault settlement, peripheral displacement, vault settlement rate, and peripheral displacement rate are composed of various measuring points as control indicators.

4. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 3 is characterized in that: The 5 levels are set up as follows: The first level, groundwater status identification interval division: poor, relatively developed, developed, abundant, extremely abundant = [0, 1], [1, 2], [2, 3], [3, 4], [4, 5]; The second level, uniaxial saturated compressive strength identification interval division: extremely soft rock, soft rock, relatively soft rock, hard rock, extremely hard rock = [0, 5], [5, 15], [15, 30], [30, 60], [60, 120]; The third level, the integrity index identification interval division: extremely broken, broken, relatively broken, relatively complete, complete = [0, 0.15], [0.15, 0.35], [0.35, 0.55], [0.55, 0.75], [0.75, 1]; At the fourth level, the identification intervals of the vault subsidence and peripheral displacement are divided into [0, 1 / 5U0], [1 / 5U0, 2 / 5U0], [2 / 5U0, 3 / 5U0], [3 / 5U0, 4 / 5U0], and [4 / 5U0, U0], where U0 is the design reserved deformation; The fifth level, the recognition interval of the vault sinking rate and the peripheral displacement rate is divided into [0, 1 / 5v0], [1 / 5v0, 2 / 5v0], [2 / 5v0, 3 / 5v0], [3 / 5v0, 4 / 5v0], [4 / 5v0, v0], where v0 is the limit rate.

5. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 4 is characterized in that: When v0 is set to 1 mm / d, the identification intervals of the dome sinking rate and the peripheral displacement rate are set as follows: [0, 0.2], [0.2, 0.4], [0.4, 0.6], [0.6, 0.8], [0.8, 1].

6. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The relative weights of the indicators described in step 4 are allocated as follows: Table of weights and support coefficients for indicators at all levels The weight distribution of the five-level indicators is evenly distributed according to the number of measuring points.

7. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The interval distance calculation formula described in step 5 is as follows: Suppose A = [a1, a2] and B = [b1, b2] are two interval numbers, then the square of their distance D 2 for: Interval distance The real number a is converted into an interval number represented as [a, a]. The distance from the original data to each identification interval is calculated according to the above formula to obtain the interval distance vector D = (d1, d2, d3, d4, d5).

8. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The calculation method of the interval similarity formula described in step 6 is as follows: Assume A = [a1, a2], B = [b1, b2] are two interval numbers, then the similarity S(A, B) of the interval numbers A and B is: Where α>0 is the support coefficient, D(A, B) is the distance between the interval numbers A and B; the support coefficient α is mainly used to adjust the discrete degree of the generated similarity value; the interval distance vector is converted into a similarity vector S=(s1, s2, s3, s4, s5) through the similarity calculation formula.

9. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1 is characterized in that: The analysis matrix described in step 7 is composed of normalized similarity vectors of multiple indicators at the same level; Assume that a certain level contains m indicators, and the weight vector of these m indicators is w = (w1, w2, ..., w m ), the normalized similarity vector of the jth index is r j =(a j1 , a j2 , a j3 , a j4 , a j5 ), j = 1, 2, ..., m. The matrix Q composed of the calculation results of the same-level indicators i for:

10. The method for dynamic prediction and early warning of construction tunnel deformation based on multi-source data fusion according to claim 1, characterized in that: The data fusion described in step 7 is calculated using the following formula: The calculation result Q of the previous level i+1 for: Q i+1 =W*Q i After layer-by-layer fusion, the security state vector (q1, q2, q3, q4, q5) of a certain section of the tunnel is finally obtained, where q1, q2, q3, q4, q5 correspond to security state levels 1, 2, 3, 4, and 5 respectively, and the security level at each time point is represented by a probability distribution of 5 levels: {1, 2, 3, 4, 5} = (q1, q2, q3, q4, q5). Among them, q1+q2+q3+q4+q5=1, and the level with the maximum probability is called the explicit state, while the other 4 security levels are in the implicit state, where the explicit state is the real security level of the tunnel.

Images