Water tunnel leakage evaluation method based on photoelectric method fusion

Abstract

The invention discloses a water conveyance tunnel leakage evaluation method based on photoelectric method fusion. The method comprises the steps of arrangement of magnetoelectric method equipment and arrangement of optical fiber sensors; acquiring data of a multi-source sensor; preprocessing the multi-source sensor data; carrying out fusion analysis on multi-source sensor data; and the leakage state of the water conveyance tunnel is evaluated by integrating the fusion data of the multi-source sensor. The monitoring results of the magnetoelectric method and the optical fiber sensing technology can be mutually verified, abnormal data are eliminated, and the monitoring results are more credible. The magnetoelectric method and the optical fiber sensing technology are fused for detection, so that the defect of a single means can be overcome, and the ultra-deep long-distance water delivery tunnel can be detected. The fusion of the magnetoelectric method data and the optical fiber data can reflect the change of multiple physical fields of the tunnel body and the surrounding rock-soil body when the water delivery tunnel is damaged and leaked, and can reflect the more real leakage state of the water delivery tunnel.

Description

Technical Field

[0001] The present invention relates to the field of structural health detection, and in particular to a water conveyance tunnel leakage evaluation method based on photoelectric method fusion. Background Art

[0002] Water transfer tunnels are an indispensable part of water resource allocation. Optical fibers are small in size and light in weight, soft and easy to bend, can adapt to various complex terrains, resist electromagnetic interference, and have low deployment costs. In addition, light wave signals have low loss when transmitted in optical fibers and have a wide bandwidth, making them suitable for long-distance detection. In underground media, electromagnetic waves have low attenuation and strong penetration ability, and are less affected by high-resistance shielding layers. Therefore, the magnetoelectric method can detect abnormal changes in deeply buried water transfer tunnels. However, the underground space environment is complex, and the monitoring results obtained by using a single method to monitor water transfer tunnel leakage have a high degree of multi-solution. Therefore, the present invention proposes a water transfer tunnel leakage evaluation method based on the fusion of optoelectronic methods. Summary of the invention

[0003] The purpose of the present invention is to provide a water conveyance tunnel leakage evaluation method based on the fusion of photoelectric method to comprehensively evaluate the leakage status of the water conveyance tunnel.

[0004] In order to achieve the above functions, the present invention designs a water tunnel leakage evaluation method based on the fusion of photoelectric method, and performs the following steps S1 to S5 for the water tunnel to complete the evaluation of the water tunnel leakage:

[0005] Step S1: deploying multi-source sensors, including deploying optical fiber sensors inside the water tunnel and within a preset range around it, and deploying magnetoelectric sensors above the water tunnel;

[0006] Step S2: multi-source sensors collect data, optical fiber sensors collect multi-physical field data of the water conveyance tunnel, and magnetoelectric sensors collect magnetic induction intensity data of the water conveyance tunnel;

[0007] Step S3: preprocessing the data collected by the multi-source sensors;

[0008] Step S4: using the unscented Kalman filter algorithm to perform fusion analysis on the multi-physical field data collected by the optical fiber sensor and the magnetic induction intensity data collected by the magnetoelectric sensor, and outputting the state estimation mean and the state estimation covariance matrix after iterative updating;

[0009] Step S5: Based on the output state estimation mean and covariance matrix, the leakage of the water transfer tunnel is evaluated by integrating the changes in multi-physical field data.

[0010] As a preferred technical solution of the present invention: the optical fiber sensor in step S1 includes a distributed optical fiber arranged along the direction of the water tunnel, an earth pressure gauge, an active heating thermometer, and a displacement meter arranged within a preset range around the water tunnel; the magnetoelectric sensor includes a magnetoelectric device arranged above the water tunnel.

[0011] As a preferred technical solution of the present invention: in step S2, the distributed optical fiber collects strain change data along the water transfer tunnel, the soil pressure gauge collects pressure change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, the active heating thermometer collects temperature data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, the soil displacement meter collects displacement change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, and the magnetoelectric method equipment collects abnormal change data of magnetic induction intensity when the water transfer tunnel leaks and is damaged.

[0012] As a preferred technical solution of the present invention: the data preprocessing in step S3 includes: data cleaning, data normalization, data spatiotemporal synchronization, and noise filtering; the data cleaning includes eliminating abnormal data and filling missing data with interpolation methods; the data normalization is to remove the dimension of the data and uniformly map it to the interval [0, 1]; the data spatiotemporal synchronization is to align all data in spatiotemporal order; the noise filtering is to smooth the data.

[0013] As a preferred technical solution of the present invention: Step S4 specifically comprises the following steps:

[0014] Step S4.1: Generate Sigma points, as follows:

[0015]

[0016] Where λ = α 2 (n+κ)-n, n is the dimension of the state vector, λ, α and κ are adjustment parameters, Sigma points extending from the mean in the positive direction, is the Sigma point extending from the mean to the negative direction, k represents the order of steps, μ k-1 is the mean of the state vector, P k-1 is the covariance matrix of the state vector, the state vector H = [A, B, C, D, E] T , A is the change of magnetic field, B is the change of temperature field, C is the change of strain field, D is the change of stress field, and E is the change of displacement field;

[0017] Step S4.2: Predict the Sigma point, as follows:

[0018]

[0019] in, is the predicted Sigma point, f is the state transfer function;

[0020] Step S4.3: Predict the state estimation mean and state estimation covariance, as shown in the following formula:

[0021]

[0022] Among them, μ k|k-1 is the predicted state mean, P k|k-1 is the predicted covariance matrix, Q is the process noise covariance matrix, and are the weights of the covariance and mean;

[0023] Step S4.4: Calculate the measurement mean and covariance, as follows:

[0024]

[0025] Among them, Z k|k-1 is the predicted observation value, h is the observation model function, P zz is the measurement covariance matrix, P xz is the cross covariance matrix of state and measurement, R is the observation noise covariance matrix;

[0026] Step S4.5: Calculate the Kalman gain and state, as follows:

[0027] K=P xz P zz -1

[0028] μ k|k =μ k|k-1 +K(zZ k|k-1 )

[0029] P k|k =P k|k-1 -KP zz K T

[0030] Among them, K is the Kalman gain, z is the actual observation value, μ k|k is the state estimated mean, P k|k is the covariance matrix.

[0031] Step S4.6: Update the state estimate mean and state estimate covariance, and use the updated mean and covariance matrix as the initial value for the next iteration. The iteration termination condition is that the change of the Euclidean norm of the state estimate and the diagonal elements of the covariance matrix is ​​less than 10 -3 , output the state estimation mean and state estimation covariance matrix at this time.

[0032] As a preferred technical solution of the present invention: the specific method for evaluating the leakage of the water transfer tunnel in step S5 is as follows:

[0033] The current state of the system is represented by the state estimation mean. If the change amplitude of the magnetic field intensity exceeds 20nT, the change amplitude of the temperature field exceeds 1℃, the change amplitude of the strain field exceeds 400με, the change amplitude of the stress field exceeds 2MPa, and the change amplitude of the displacement field exceeds 5cm, the water diversion tunnel is considered to have a leakage risk. The diagonal elements of the covariance matrix represent the variance of each variable in the state estimation mean. If the variance of the magnetic field intensity is less than 4, the variance of the temperature field is less than 0.04, the variance of the strain field is less than 6400, the variance of the stress field is less than 0.16, and the variance of the displacement field is less than 1, the state estimation mean is considered to have high quality and small uncertainty.

[0034] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0035] 1. Magnetoelectric method and fiber optic sensing technology can verify the monitoring results with each other, eliminate abnormal data, and make the monitoring results more credible.

[0036] 2. The fusion detection of magnetoelectric method and fiber optic sensing technology can overcome the shortcomings of a single method and realize the detection of ultra-deep and long-distance water tunnels.

[0037] 3. The fusion of magneto-electric data and optical fiber data can reflect the changes in the multi-physical fields of the tunnel body and the surrounding rock and soil when the water transfer tunnel is damaged and leaking, and can reflect a more realistic leakage state. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of a water transfer tunnel leakage evaluation method based on photoelectric method fusion provided according to an embodiment of the present invention;

[0039] Figure 2 is a flow chart of a data fusion method provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0041] The embodiment of the present invention provides a method for evaluating water tunnel leakage based on photoelectric method fusion, referring to Figure 1 , perform the following steps S1 to S5 for the water diversion tunnel to complete the evaluation of the water diversion tunnel leakage:

[0042] Step S1: deploying multi-source sensors, including deploying optical fiber sensors inside the water tunnel and within a preset range around it, and deploying magnetoelectric sensors above the water tunnel;

[0043] The fiber optic sensor includes distributed optical fiber laid along the direction of the water tunnel, and earth pressure gauges, active heating thermometers, and displacement meters laid within a preset range around the water tunnel; the magnetoelectric sensor includes magnetoelectric equipment arranged above the water tunnel.

[0044] The magnetoelectric method can use the m-sequence pseudo-random algorithm to enhance the noise resistance and detection accuracy of the magnetoelectric method. The magnetoelectric method equipment can use drones, unmanned vehicles, unmanned ships and other vehicles to meet different usage scenarios.

[0045] Step S2: multi-source sensors collect data, optical fiber sensors collect multi-physical field data of the water conveyance tunnel, and magnetoelectric sensors collect magnetic induction intensity data of the water conveyance tunnel;

[0046] Unmanned sites are built in the field, and magnetoelectric equipment and fiber optic equipment continuously collect data. The collected data is uploaded to the cloud. Researchers log in to the cloud to receive the data, realizing long-term unmanned detection.

[0047] Distributed optical fiber collects strain change data along the water transfer tunnel, soil pressure gauge collects pressure change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, and active heating thermometer collects temperature data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, so as to monitor the impact of water transfer tunnel leakage and damage on the moisture content of the surrounding rock and soil; soil displacement meter collects displacement change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, and magnetoelectric equipment collects abnormal change data of magnetic induction intensity when the water transfer tunnel leaks and is damaged.

[0048] Step S3: preprocessing the data collected by the multi-source sensors;

[0049] Data preprocessing includes: data cleaning, data normalization, data spatiotemporal synchronization, and noise filtering; the data cleaning includes removing abnormal data and filling missing data with interpolation methods; the data normalization is to remove the dimension of the data and uniformly map it to the interval [0, 1]; the data spatiotemporal synchronization is to align all data in spatiotemporal order; the noise filtering is to smooth the data.

[0050] Step S4: using the unscented Kalman filter algorithm to perform fusion analysis on the multi-physical field data collected by the optical fiber sensor and the magnetic induction intensity data collected by the magnetoelectric sensor, and outputting the state estimation mean and the state estimation covariance matrix after iterative updating;

[0051] Reference Figure 2 , the specific steps of step S4 are as follows:

[0052] Step S4.1: Generate Sigma points, as follows:

[0053]

[0054] Where λ = α 2 (n+κ)-n, n is the dimension of the state vector, λ, α and κ are adjustment parameters, Sigma points extending from the mean in the positive direction, is the Sigma point extending from the mean to the negative direction, k represents the order of steps, μ k-1 is the mean of the state vector, P k-1 is the covariance matrix of the state vector, the state vector H = [A, B, C, D, E] T , A is the change of magnetic field, B is the change of temperature field, C is the change of strain field, D is the change of stress field, and E is the change of displacement field;

[0055] Step S4.2: Predict the Sigma point, as follows:

[0056]

[0057] in, is the predicted Sigma point, f is the state transfer function;

[0058] Step S4.3: Predict the state estimation mean and state estimation covariance, as shown in the following formula:

[0059]

[0060] Among them, μ k|k-1 is the predicted state mean, P k|k-1 is the predicted covariance matrix, Q is the process noise covariance matrix, and are the weights of the covariance and mean;

[0061] Step S4.4: Calculate the measurement mean and covariance, as follows:

[0062]

[0063] Among them, Z k|k-1 is the predicted observation value, h is the observation model function, P zz is the measurement covariance matrix, P xz is the cross covariance matrix of state and measurement, R is the observation noise covariance matrix;

[0064] Step S4.5: Calculate the Kalman gain and state, as follows:

[0065] K=P xz P zz -1

[0066] μ k|k=μ k|k-1 +K(zZ k|k-1 )

[0067] P k|k =P k|k-1 -KP zz K T

[0068] Among them, K is the Kalman gain, z is the actual observation value, μ k|k is the state estimated mean, P k|k is the covariance matrix.

[0069] Step S4.6: Update the state estimate mean and state estimate covariance, and use the updated mean and covariance matrix as the initial value for the next iteration. The iteration termination condition is that the change of the Euclidean norm of the state estimate and the diagonal elements of the covariance matrix is ​​less than 10 -3 , output the state estimation mean and state estimation covariance matrix at this time.

[0070] Step S5: For the output state estimation mean and covariance matrix, the leakage of the water tunnel is evaluated by integrating the changes in multi-physical field data. The state estimation mean is the best estimate of the current state of the system. When the change amplitude of the magnetic field intensity exceeds 20nT, the change amplitude of the temperature field exceeds 1℃, the change amplitude of the strain field exceeds 400με, the change amplitude of the stress field exceeds 2MPa, and the change amplitude of the displacement field exceeds 5cm, it can be considered that the water tunnel has a large leakage risk and needs long-term close dynamic monitoring. The diagonal elements of the covariance matrix represent the variance of each variable in the state estimation mean. When the variance of the magnetic field intensity is less than 4, the variance of the temperature field is less than 0.04, the variance of the strain field is less than 6400, the variance of the stress field is less than 0.16, and the variance of the displacement field is less than 1, it can be considered that the state estimation mean has high quality and low uncertainty.

[0071] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A water tunnel leakage evaluation method based on photoelectric method fusion, characterized in that: The following steps S1 to S5 are performed for the water conveyance tunnel to complete the evaluation of the leakage of the water conveyance tunnel: Step S1: deploying multi-source sensors, including deploying optical fiber sensors inside the water tunnel and within a preset range around it, and deploying magnetoelectric sensors above the water tunnel; Step S2: multi-source sensors collect data, optical fiber sensors collect multi-physical field data of the water conveyance tunnel, and magnetoelectric sensors collect magnetic induction intensity data of the water conveyance tunnel; Step S3: preprocessing the data collected by the multi-source sensors; Step S4: using the unscented Kalman filter algorithm to perform fusion analysis on the multi-physical field data collected by the optical fiber sensor and the magnetic induction intensity data collected by the magnetoelectric sensor, and outputting the state estimation mean and the state estimation covariance matrix after iterative updating; Step S5: Based on the output state estimation mean and covariance matrix, the leakage of the water transfer tunnel is evaluated by integrating the changes in multi-physical field data.

2. The method for evaluating water tunnel leakage based on the fusion of photoelectric method according to claim 1 is characterized in that: The optical fiber sensor in step S1 includes a distributed optical fiber arranged along the direction of the water tunnel, and an earth pressure gauge, an active heating thermometer, and a displacement meter arranged within a preset range around the water tunnel; the magnetoelectric sensor includes a magnetoelectric device arranged above the water tunnel.

3. The method for evaluating water tunnel leakage based on the fusion of photoelectric method according to claim 2 is characterized in that: In step S2, the distributed optical fiber collects strain change data along the water transfer tunnel, the soil pressure gauge collects pressure change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, the active heating thermometer collects temperature data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, the soil displacement meter collects displacement change data of the surrounding rock and soil when the water transfer tunnel leaks and is damaged, and the magnetoelectric method equipment collects abnormal change data of magnetic induction intensity when the water transfer tunnel leaks and is damaged.

4. The method for evaluating water tunnel leakage based on photoelectric method fusion according to claim 1 is characterized in that: The data preprocessing in step S3 includes: data cleaning, data normalization, data spatiotemporal synchronization, and noise filtering; the data cleaning includes removing abnormal data and filling missing data with interpolation methods; the data normalization is to remove the dimension of the data and uniformly map it to the interval [0, 1]; the data spatiotemporal synchronization is to align all data in spatiotemporal order; the noise filtering is to smooth the data.

5. The method for evaluating water tunnel leakage based on photoelectric method fusion according to claim 1 is characterized in that: The specific steps of step S4 are as follows: Step S4.1: Generate Sigma points, as follows: Where λ = α 2 (n+κ)-n, n is the dimension of the state vector, λ, α and κ are adjustment parameters, Sigma points extending from the mean in the positive direction, is the Sigma point extending from the mean to the negative direction, k represents the order of steps, μ k-1 is the mean of the state vector, P k-1 is the covariance matrix of the state vector, the state vector H = [A, B, C, D, E] T , A is the change of magnetic field, B is the change of temperature field, C is the change of strain field, D is the change of stress field, and E is the change of displacement field; Step S4.2: Predict the Sigma point, as follows: in, is the predicted Sigma point, f is the state transfer function; Step S4.3: Predict the state estimation mean and state estimation covariance, as shown in the following formula: Among them, μ k|k-1 is the predicted state mean, P k|k-1 is the predicted covariance matrix, Q is the process noise covariance matrix, and are the weights of the covariance and mean; Step S4.4: Calculate the measurement mean and covariance, as follows: Among them, Z k|k-1 is the predicted observation value, h is the observation model function, P zz is the measurement covariance matrix, P xz is the cross covariance matrix of state and measurement, R is the observation noise covariance matrix; Step S4.5: Calculate the Kalman gain and state, as follows: K=P xz P zz -1 μ k|k =μ k|k-1 +K(zZ k|k-1 ) P k|k =P k|k-1 -KP zz K T Among them, K is the Kalman gain, z is the actual observation value, μ k|k is the state estimated mean, P k|k is the covariance matrix. Step S4.6: Update the state estimate mean and state estimate covariance, and use the updated mean and covariance matrix as the initial value for the next iteration. The iteration termination condition is that the change of the Euclidean norm of the state estimate and the diagonal elements of the covariance matrix is ​​less than 10 -3 , output the state estimation mean and state estimation covariance matrix at this time.

6. The method for evaluating water tunnel leakage based on photoelectric method fusion according to claim 1 is characterized in that: The specific method for evaluating the leakage of the water transfer tunnel in step S5 is as follows: The current state of the system is represented by the state estimation mean. If the change amplitude of the magnetic field intensity exceeds 20nT, the change amplitude of the temperature field exceeds 1℃, the change amplitude of the strain field exceeds 400με, the change amplitude of the stress field exceeds 2MPa, and the change amplitude of the displacement field exceeds 5cm, the water diversion tunnel is considered to have a leakage risk. The diagonal elements of the covariance matrix represent the variance of each variable in the state estimation mean. If the variance of the magnetic field intensity is less than 4, the variance of the temperature field is less than 0.04, the variance of the strain field is less than 6400, the variance of the stress field is less than 0.16, and the variance of the displacement field is less than 1, the state estimation mean is considered to have high quality and small uncertainty.

Images