Sensor Layout Optimization Method Based on Information Entropy and Mutual Information Redundancy Elimination Principle

Abstract

The present invention provides a sensor layout optimization method based on the principle of information entropy and mutual information redundancy elimination, including: Step 1, dividing the entire cylinder cabin area into grids, binning according to the temperature difference span, calculating the temperature information entropy of the cylinder cabin area, and quantifying the temperature probability distribution characteristics through information entropy; Step 2, optimizing the optical fiber sensing path based on the Bayesian algorithm, setting a rectangular optical fiber sensing path for the cylindrical structure, with the optimization goal of a large sum of information entropy and a small length of the optical fiber sensing path, and the perturbation method being random perturbation of the rectangular vertices; Step 3, calculating the mutual information between adjacent sensors inside the optical fiber sensing path, and if the mutual information of consecutive measurement points is greater than the set redundancy threshold, identifying and removing redundant sensors to further optimize the number of sensors inside the optical fiber sensing path. The present invention optimizes the distributed optical fiber sensing path, reduces the arrangement of unnecessary measurement points, and reduces the use cost and failure risk on the premise of having little impact on the temperature field reconstruction accuracy.

Description

Technical Field

[0001] The present invention belongs to the field of temperature monitoring layout optimization based on distributed optical fiber sensors, and particularly relates to a sensor layout optimization method based on the principle of information entropy and mutual information redundancy elimination. Background Art

[0002] Spacecraft face complex thermal environments in space. Instrumentation and equipment in the cylinder compartment, such as attitude control systems, data transmission devices, etc., require stable temperature conditions. Temperature anomalies may cause the operating parameters of the equipment to drift, affecting the attitude control accuracy of the spacecraft and the accuracy of data transmission. For example, when the temperature of optical instruments on a satellite changes significantly, lens distortion may occur, affecting the imaging quality and thus the observation and monitoring of the Earth.

[0003] In the current fields of engineering applications and scientific research, distributed optical fiber sensors, with their unique advantage of being able to achieve long-distance and continuous monitoring of physical quantities such as temperature and strain, have shown great application potential in many scenarios such as complex temperature field monitoring. Currently, the commonly used distributed optical fiber sensor layout optimization methods mainly include neural network methods, particle swarm optimization methods, etc. However, when dealing with the monitoring requirements of complex temperature fields, these methods have exposed a series of problems that need to be solved urgently.

[0004] The neural network method is highly dependent on data. Especially for the monitoring of complex temperature fields, it makes the data acquisition process full of uncertainties and it is difficult to obtain sufficient effective samples. The interpretability of the neural network model is poor. In the monitoring of complex temperature fields, it is impossible to understand the design basis and internal logic of the scheme, which limits its application in actual engineering. In addition, the training process of the neural network involves a large amount of matrix operations and iterative optimization, with extremely high computational complexity, and it requires a large amount of computing resources and time.

[0005] The particle swarm optimization method is prone to falling into local optimal solutions and unable to continue exploring the global optimal solution. Secondly, as the number of iterations increases, the convergence speed of the particle swarm optimization algorithm will slow down significantly. In addition, the particle swarm optimization method lacks an effective constraint handling mechanism. In the actual monitoring of complex temperature fields, the layout of sensors is usually restricted by various constraint conditions, such as the number limit of sensors, physical limitations of the layout space, cost limitations, etc. The particle swarm algorithm itself does not have a built-in effective method to handle these constraint conditions and requires additional complex technologies for constraint handling. Summary of the Invention

[0006] Object of the Invention: The technical problem to be solved by the present invention is to provide a sensor layout optimization method based on the principle of information entropy and mutual information redundancy elimination for the deficiencies of the prior art, including the following steps:

[0007] Step 1: Divide the entire cylinder cabin area into grids, bin them according to the temperature difference span, calculate the temperature information entropy of the cylinder cabin area, and quantify the temperature probability distribution characteristics through the information entropy;

[0008] Step 2: Optimize the optical fiber sensing path based on the Bayesian algorithm. Set a rectangular optical fiber sensing path for the cylindrical structure. The optimization goal is to maximize the total information entropy and minimize the length of the optical fiber sensing path. The perturbation method is random perturbation of the rectangular vertices;

[0009] Step 3: For the optimized optical fiber sensing path obtained in Step 2, calculate the mutual information between adjacent sensors inside the optical fiber sensing path. If the mutual information of consecutive measurement points is greater than the set redundancy threshold, identify and remove redundant sensors to further optimize the number of sensors inside the optical fiber sensing path.

[0010] Step 1 includes:

[0011] Step 1-1: For the problem of calculating the temperature information entropy of the cylinder cabin area, divide the entire cylinder cabin area into m×n grids. The grid in the i-th row and j-th column is denoted as G ij , where m is the total number of rows, n is the total number of columns, i ranges from 1 to m, and j ranges from 1 to n. Assume there are k temperature measurement points {T ij , T ij,1 , …, T ij,2 , …, T ij,k} in the grid G min , T max . Bin the entire temperature range [T

[0012] (1)

[0013] according to the temperature difference span ΔT into L temperature intervals: ij,k represents that there are k temperature measurement points in total in the grid G ij ; T min represents the minimum temperature in the grid G ij , and T max represents the maximum temperature in the grid G ij , which are obtained by querying the finite element simulation results;

[0014] Step 1-2: Inside the grid G ij , assume the l -th temperature interval is T l , T l+1 , where T l is the lower temperature limit of the l -th temperature interval, and T l+1 is the lThe upper temperature limit of a temperature range, within the grid G ij Inside, the temperature value of the temperature measurement point is in the l probability P of the temperature range ij,l is the number of times corresponding to the l th temperature range divided by the total number k of temperature measurement points:

[0015] (2),

[0016] where represents the number of times corresponding to the l th temperature range among k temperature measurement points;

[0017] The temperature information entropy H(G ij ) of the grid G is calculated through the following formula: ij )

[0018] (3).

[0019] Step 2 includes: setting a rectangular optical fiber sensing path for the cylindrical structure. The four vertex coordinates of the rectangular optical fiber sensing path are (a1, b1), (a1, b2), (a2, b1), and (a2, b2) respectively. The optimization objective of the rectangular optical fiber sensing path is that the total information entropy of the sensing path coverage area is large and the sensor laying path length is small. A large total information entropy of the sensing path coverage area means a large uncertainty in the path temperature information, and a small sensing path length means less sensor usage. The optimization objective function is:

[0020] (4),

[0021] where f(Path) represents the sensing path optimization function, Path represents the sensing path, H path is the total information entropy of the area covered by the sensing path, L path is the total length of the sensing path, and α and β are weight coefficients.

[0022] Step 2 also includes: setting the set of grids passed by the rectangular optical fiber sensing path as , and the information entropies of the x grids passed by the rectangular optical fiber sensing path in sequence are respectively, where represents the information entropy of the pth grid on the rectangular optical fiber sensing path, then the total information entropy of all the grids passed by the rectangular optical fiber sensing path is:

[0023] (5),

[0024] where x is the total number of grids passed by the rectangular optical fiber sensing path.

[0025] Step 2 further includes: the total length L of the rectangular optical fiber sensing path path The calculation formula is:

[0026] (6),

[0027] where b p is the width of the p-th information entropy grid passed by the rectangular optical fiber sensing path;

[0028] If it is necessary to control the change of the rectangular optical fiber sensing path, it is only necessary to control the change of the diagonal vertex coordinates (a1, b1) and (a2, b2) of the rectangular optical fiber sensing path. The perturbation function is set as:

[0029] (7),

[0030] where represents a randomly selected integer greater than or equal to -t and less than or equal to t, representing the perturbation of the diagonal vertex coordinate a1 of the rectangular optical fiber sensing path, controlling the change of a1, and t is a random integer;

[0031] represents a randomly selected integer greater than or equal to -t and less than or equal to t, representing the perturbation of the diagonal vertex coordinate a2 of the rectangular optical fiber sensing path, controlling the change of a2;

[0032] represents a randomly selected integer greater than or equal to -t and less than or equal to t, representing the perturbation of the diagonal vertex coordinate b1 of the rectangular optical fiber sensing path, controlling the change of b1;

[0033] represents a randomly selected integer greater than or equal to -t and less than or equal to t, representing the perturbation of the diagonal vertex coordinate b2 of the rectangular optical fiber sensing path, controlling the change of b2.

[0034] Step 2 further includes: when moving the vertices of the rectangular optical fiber sensing path, it is necessary to control the diagonal vertex coordinates (a1, b1) and (a2, b2) within the monitoring range, and the following formula needs to be satisfied:

[0035] (8),

[0036] where, represents the length of the cylinder cabin monitoring area, and b represents the width of the cylinder cabin monitoring area.

[0037] Step 3 includes:

[0038] Calculate the mutual information between two adjacent grids G i,j and G i+1,j passed by the rectangular optical fiber sensing path. Each of the two adjacent grids contains k temperature measurement points. The two adjacent grids Gi,j and G i+1,j The sets of temperature values are T i,j and T i+1,j ;

[0039] T i,j = {T i,j,1 , T i,j,2 , …, T i,j,K},

[0040] T i+1,j = {T i+1,j,1 , T i+1,j,2 , …, T i+1,j,K},

[0041] where T i,j,K is the temperature value of the k-th temperature measurement point in grid G i,j , and T i+1,j,K is the temperature value of the k-th temperature measurement point in grid G i+1,j .

[0042] Step 3 further includes:

[0043] Equal-width binning is performed on the two sets of grid temperature values T i,j and T i+1,j simultaneously, divided into L regions according to the temperature difference span ΔT, and the number of occurrences of specific temperature values in each temperature interval is counted. Let the c-th temperature interval be [T c , T c+1 , and let the d-th temperature interval be [T d , T d+1 , where T c represents the lower temperature limit of the c-th temperature interval, T c+1 represents the upper temperature limit of the c-th temperature interval, T d represents the lower temperature limit of the d-th temperature interval, and T d+1 represents the upper temperature limit of the d-th temperature interval;

[0044] In grid G i,j , the number of times corresponding to the c-th temperature interval [T c , T c+1 is C i,j,c ;

[0045] In grid G i+1,j , the number of times corresponding to the d-th temperature interval [T d , T d+1 is C i+1,j,d ;

[0046] For each interval combination (c, d), count the two adjacent grids G i,j and G i+1,jThe number C of times when the temperature values at the corresponding positions fall within the intervals c and d respectively c,d , calculate the marginal probabilities P i,j and P i+1,j of two adjacent grids G i,j (G p ) and P i+1,j (G q ), as well as the joint probability distribution P(G p ,G q ):

[0047] , , (9),

[0048] where represents the number of times when the c-th temperature interval within the grid G i+1,j is [T d ,T d+1 ;

[0049] The mutual information I(T i,j and G i+1,j ) between two adjacent grids G i,j and G i+1,j is:[[]]

[0050] (10),

[0051] The mutual information itself has no fixed upper limit, which makes it difficult to compare the mutual information between different data sets. Normalize the mutual information of adjacent points on the rectangular optical fiber sensing path. The formula is:

[0052] (11),

[0053] where NMI(T i,j and T i+1,j ) is the normalized mutual information, and min{H(T i,j ), H(T i+1,j )} is the smaller value of the temperature information entropy of two adjacent grids G i,j and G i+1,j , which is calculated by formula (3);

[0054] After completing the normalization calculation of the mutual information of all adjacent measurement points on the rectangular optical fiber sensing path, set the redundancy threshold NMI threshold . Generally, when NMI is greater than 0.4, it can be judged that there is an obvious correlation. Let NMI threshold =0.4. When the mutual information of consecutive measurement points is greater than this threshold NMI threshold , then the sensor is judged as a redundant sensor.

[0055] The present invention also provides an electronic device, including a processor and a memory. The memory stores program codes. When the program codes are executed by the processor, the processor is caused to execute the steps of the method described above.

[0056] The present invention also provides a storage medium storing a computer program or instruction. When the computer program or instruction runs on a computer, the steps of the method described above are executed.

[0057] Beneficial effects: The method of the present invention reduces the arrangement of unnecessary measuring points, and reduces the use cost and failure risk on the premise of having little influence on the reconstruction accuracy of the temperature field. A temperature gradient evaluation method based on the information entropy theory is proposed, combined with the Bayesian optimization method, which improves the sensitive characteristics of temperature gradient perception. A method for redundant elimination of rectangular optical fiber sensing paths based on mutual information is developed, realizing the layout optimization of distributed optical fiber sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The following further describes the present invention in detail with reference to the drawings and specific embodiments, and the above and / or other advantages of the present invention will become clearer.

[0059] Figure 1 It is a schematic diagram of the path distribution of an OFDR distributed optical fiber sensor.

[0060] Figure 2 It is a flowchart of the present invention.

[0061] Figure 3 Information entropy calculation result.

[0062] Figure 4 Changes in the rectangular optical fiber sensing path after Bayesian optimization.

[0063] Figure 5 Calculation result of the mutual information of the rectangular optical fiber sensing path passing through adjacent grids.

[0064] Figure 6 Constructing an optical fiber sensing path by deleting redundant paths based on mutual information. SPECIFIC EMBODIMENTS

[0065] As Figure 2 shown, the embodiment of the present invention provides a sensor layout optimization method based on the principles of information entropy and mutual information redundancy elimination, including:

[0066] Step 1, dividing the entire cylinder cabin area into grids, binning according to the temperature difference span, calculating the temperature information entropy of the cylinder cabin area, and quantifying the temperature probability distribution characteristics through the information entropy; including:

[0067] Step 1-1: Simulate the temperature field distribution characteristics of the cylinder cabin structure under this thermal load condition by means of finite element simulation. For the problem of calculating the temperature information entropy in the cylinder cabin area, unfold the entire cylindrical shell of the cylinder cabin into a plane, divide the entire plane area into m×n grids, which are respectively represented as G ij , where i and j represent the row number and column number. There are k temperature measurement points in each unit grid, and the temperature information of each temperature measurement point is collected, which is represented as {T ij,1 , T ij,2 ,…, T ij,k}. The entire temperature range [T min , T max is statistically analyzed, binned according to the temperature difference span ΔT, and divided into L temperature intervals, .

[0068] Step 1-2: Count the number of temperature measurement points contained in each temperature interval. For example, the l-th temperature interval is [T l , T l+1 , and its corresponding number is G ij,l . Then the probability P ij,l of this interval is . Calculate the temperature information entropy H(G ij ): ij . Represent the information entropy size with different colors, as shown in . Figure 3 .

[0069] Step 2: Optimize the optical fiber sensing path based on the Bayesian algorithm. Set a rectangular optical fiber sensing path for the cylindrical structure. The optimization goal is to maximize the sum of information entropy and minimize the length of the optical fiber sensing path. The perturbation method is random perturbation of the rectangular vertices;

[0070] Set the coordinates of the four vertices of the rectangular optical fiber sensing path as (a1, b1), (a1, b2), (a2, b1), and (a2, b2) respectively, as shown in Figure 1 . The set of grids passed by the rectangular optical fiber sensing path is . The information entropies of the x grids passed by the rectangular optical fiber sensing path in sequence are , where represents the information entropy of the p-th grid on the rectangular optical fiber sensing path. Then the sum of the information entropies of all the grids passed by the rectangular optical fiber sensing path is . Among them, x is the total number of grids passed by the rectangular optical fiber sensing path; the total length L path of the rectangular optical fiber sensing path is the sum of the grid widths passed by the rectangular optical fiber sensing path .

[0071] Control the rectangular optical fiber to sense path changes, and control the diagonal vertex coordinates (a1, b1) and (a2, b2) of the rectangular optical fiber sensing path to change randomly. Set the perturbation function as:

[0072] ,

[0073] where represents randomly selecting an integer greater than or equal to -t and less than or equal to t, which is the perturbation of the diagonal vertex coordinate a1 of the rectangular optical fiber sensing path to control the change of a1, and t is a random integer;

[0074] represents randomly selecting an integer greater than or equal to -t and less than or equal to t, which is the perturbation of the diagonal vertex coordinate a2 of the rectangular optical fiber sensing path to control the change of a2;

[0075] represents randomly selecting an integer greater than or equal to -t and less than or equal to t, which is the perturbation of the diagonal vertex coordinate b1 of the rectangular optical fiber sensing path to control the change of b1;

[0076] represents randomly selecting an integer greater than or equal to -t and less than or equal to t, which is the perturbation of the diagonal vertex coordinate b2 of the rectangular optical fiber sensing path to control the change of b2.

[0077] When moving the vertices of the rectangular optical fiber sensing path, control the diagonal vertex coordinates (a1, b1) and (a2, b2) within the monitoring range. As Figure 1 shown, set the boundary control conditions . After Bayesian optimization, the change of the rectangular optical fiber sensing path is as Figure 4 shown.

[0078] Step 3: For the optimized optical fiber sensing path obtained in Step 2, calculate the mutual information between adjacent sensors inside the optical fiber sensing path. If the mutual information of consecutive measurement points is greater than the set redundancy threshold, identify and remove redundant sensors to further optimize the number of sensors inside the optical fiber sensing path.

[0079] Calculate the mutual information between all adjacent grids G i,j and G i+1,j in the grid passed by the rectangular optical fiber sensing path. Each of the two adjacent grids contains k temperature measurement points. The temperature value sets of the two adjacent grids G i,j and G i+1,j are respectively: T i,j = {T i,j,1 , T i,j,2 , …, T i,j,K} and T i+1,j = {T i+1,j,1 , T i+1,j,2 , …, Ti+1,j,K}, T i,j,K is the temperature value of the k-th temperature measurement point in the grid G i,j , and T i+1,j,K is the temperature value of the k-th temperature measurement point in the grid G i+1,j ;

[0080] Bin the two sets of grid temperature values T i,j and T i+1,j simultaneously with equal-width binning. Divide them into L regions according to the temperature difference span ΔT, and count the number of occurrences of specific temperature values in each temperature interval. Let the c-th temperature interval be [T c , T c+1 , and let the d-th temperature interval be [T d , T d+1 . In the grid G i,j , the number of times corresponding to the c-th temperature interval [T c , T c+1 is C i,j,c . In the grid G i+1,j , the number of times corresponding to the d-th temperature interval [T d , T d+1 is C i+1,j,d ;

[0081] For each interval combination (c, d), count the number of times C i,j that the temperature values at the corresponding positions in two adjacent grids G i+1,j and G c,d fall into interval c and interval d respectively. Calculate the marginal probabilities and of two adjacent grids G i,j and G i+1,j , as well as the joint probability distribution . Calculate the mutual information of two adjacent grids G i,j and G i+1,j . The calculation results are as Figure 5 shown.

[0082] Normalize the mutual information of adjacent points on the rectangular optical fiber sensing path. Divide the mutual information I(T i,j , T i+1,j ) of two adjacent grids G i,j , T i+1,j ) by the smaller value min{H(T i,j ), H(T i+1,j )} of the temperature information entropy of adjacent grids, . Set the redundancy threshold NMI of the mutual information threshold= 0.4. When the mutual information of consecutive measurement points is greater than 0.4, the sensor is determined to be a redundant sensor, and this part of the path is deleted. The remaining part is the final optimized result of the rectangular fiber optic sensing path, as shown by the red straight line in Figure 6 as shown.

[0083] The present invention provides a method for optimizing the layout of sensors based on the principle of information entropy and mutual information redundancy elimination. There are many methods and ways to specifically implement this technical solution. The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention. Each component not clearly defined in this embodiment can be implemented using existing technologies.

Claims

1. A sensor layout optimization method based on the principle of information entropy and mutual information redundancy elimination, characterized in that It includes the following steps: Step 1: Divide the entire cylindrical cabin area into grids, bin them according to the temperature difference span, calculate the temperature information entropy of the cylindrical cabin area, and quantify the temperature probability distribution characteristics through the information entropy. Step 2: Optimize the optical fiber sensing path based on the Bayesian algorithm. Set a rectangular optical fiber sensing path for the cylindrical structure. The optimization goal is to maximize the total information entropy and minimize the length of the optical fiber sensing path. The perturbation method is to randomly perturb the rectangular vertices. Step 3: For the optimized optical fiber sensing path obtained in Step 2, calculate the mutual information between adjacent sensors inside the optical fiber sensing path. If the mutual information of consecutive measurement points is greater than the set redundancy threshold, identify and remove redundant sensors to further optimize the number of sensors inside the optical fiber sensing path.

2. The method according to claim 1, wherein Step 1 includes: Step 1-1. For the problem of calculating the temperature information entropy of the silo area, divide the entire silo area into m×n grids. The grid in the i-th row and j-th column is denoted as G ij , where m is the total number of rows, n is the total number of columns, i ranges from 1 to m, and j ranges from 1 to n. Assume that there are k temperature measurement points {T ij , T ij,1 , …, T ij,2} in the grid G ij,k . Divide the entire temperature range [T min , T max into L temperature intervals according to the temperature difference span ΔT: (1)， where T ij,k represents that there are k temperature measurement points in the grid G ij ; T min represents the minimum temperature in the grid G ij ; T max represents the maximum temperature in the grid G ij , which is obtained by querying the finite element simulation results; Step 1-2, within the grid G ij Inside, let the l th temperature range be T l , T l+1 , where T l is the lower temperature limit of the l th temperature range, and T l+1 is the upper temperature limit of the l th temperature range. Inside the grid G ij , the probability P l that the temperature measurement point temperature value is within the ij,l th temperature range is the number of times corresponding to the l th temperature range divided by the total number k of temperature measurement points: (2)， wherein represents the number of times corresponding to the l th temperature range among k temperature measurement points; Calculate the temperature information entropy \(H(G)\) of the grid \(G\) through the following formula ij : ij ) (3)。 3. The method according to claim 2, wherein Step 2 includes: Set a rectangular optical fiber sensing path for the cylindrical structure. The four vertex coordinates of the rectangular optical fiber sensing path are (a1, b1), (a1, b2), (a2, b1), and (a2, b2) respectively. The optimization goal of the rectangular optical fiber sensing path is to maximize the total information entropy of the area covered by the sensing path and minimize the length of the sensor laying path. The optimization objective function is: (4)， where f(Path) represents the perception path optimization function, H path is the total information entropy of the area covered by the perception path, L path is the total length of the perception path, and α and β are weight coefficients.

4. The method according to claim 3, wherein Step 2 further includes: setting the set of grids through which the rectangular optical fiber sensing path passes as , and the information entropy of x grids successively passed by the rectangular optical fiber sensing path are respectively , where represents the p-th grid on the rectangular optical fiber sensing path , then the total information entropy of all the grids passed by the rectangular optical fiber sensing path is: (5)， where x is the total number of grids passed by the rectangular optical fiber sensing path.

5. The method according to claim 4, wherein Step 2 further includes: the total length L of the rectangular optical fiber sensing path path The calculation formula is as follows: (6)， where b p is the width of the p-th information entropy grid passed by the rectangular optical fiber sensing path; If it is necessary to control the change of the rectangular optical fiber sensing path, only need to control the change of the diagonal vertex coordinates (a1, b1) and (a2, b2) of the rectangular optical fiber sensing path. Set the perturbation function as: (7)， Among them represents randomly selecting an integer greater than or equal to -t and less than or equal to t to perturb the coordinates a1 of the diagonal vertices of the rectangular optical fiber sensing path, controlling the change of a1, and t is a random integer; It means randomly selecting an integer greater than or equal to -t and less than or equal to t to perturb the coordinates of the diagonal vertices a2 of the rectangular optical fiber sensing path and control the change of a2; It means randomly selecting an integer greater than or equal to -t and less than or equal to t to perturb the coordinates of the diagonal vertices b1 of the rectangular optical fiber sensing path and control the change of b1; It means randomly selecting an integer greater than or equal to -t and less than or equal to t to perturb the coordinates of the diagonal vertices b2 of the rectangular optical fiber sensing path and control the change of b2.

6. The method according to claim 5, characterized in that, Step 2 also includes: When moving the vertices of the rectangular optical fiber sensing path, it is necessary to control the diagonal vertex coordinates (a1, b1) and (a2, b2) within the monitoring range, and the following formula needs to be satisfied: (8)， Among them, represents the length of the monitoring area of the cylinder cabin, and b represents the width of the monitoring area of the cylinder cabin.

7. The method according to claim 6, characterized in that, Step 3 includes: Calculate the mutual information between two adjacent grids G i,j and G i+1,j through which the rectangular optical fiber sensing path passes. Each of the two adjacent grids contains k temperature measurement points. The temperature value sets of the two adjacent grids G i,j and G i+1,j are T i,j and T i+1,j respectively; T i,j = {T i,j,1 , T i,j,2 ,…, T i,j,K}, T i+1,j = {T i+1,j,1 , T i+1,j,2 , …, T i+1,j,K}, where T i,j,K is the temperature value of the k-th temperature measurement point in the grid G i,j , and T i+1,j,K is the temperature value of the k-th temperature measurement point in the grid G i+1,j .

8. The method according to claim 7, wherein Step 3 also includes: Two sets of grid temperature values T i,j and T i+1,j are simultaneously binned with equal width and divided into L regions according to the temperature difference span ΔT. The number of occurrences of specific temperature values in each temperature range is counted. Let the c-th temperature range be [T c , T c+1 , and let the d-th temperature range be [T d , T d+1 , where T c represents the lower temperature limit of the c-th temperature range, T c+1 represents the upper temperature limit of the c-th temperature range, T d represents the lower temperature limit of the d-th temperature range, and T d+1 represents the upper temperature limit of the d-th temperature range; Within grid G i,j the number of times corresponding to the c-th temperature range [T c , T c+1 is C i,j,c ; Within grid G i+1,j the d-th temperature range is [T d , T d+1 and the corresponding number of times is C i+1,j,d ; For each interval combination (c, d), count the two adjacent grids G i,j and G i+1,j The number of times the temperature value at the corresponding position falls in interval c and interval d respectively is C c,d , calculate two adjacent grids G respectively i,j and G i+1,j The marginal probability P i,j (G p ) and P i+1,j (G q ), and the joint probability distribution P(G p ,G q ): ， , (9)， wherein represents the number of times corresponding to the c-th temperature range [T i+1,j , T d , T d+1 within the grid G; The mutual information I(T i,j , T i+1,j ) between two adjacent meshes G i,j and G i+1,j is as follows: (10)， Normalize the mutual information between adjacent points of the rectangular optical fiber sensing path. The formula is: (11)， where NMI(T i,j , T i+1,j ) is the normalized mutual information, and min{H(T i,j ), H(T i+1,j )} is the smaller value of the temperature information entropy of two adjacent grids G i,j and G i+1,j , which is calculated by formula (3); After completing the normalized calculation of the mutual information for all adjacent measurement points on the rectangular optical fiber sensing path, set the redundancy threshold NMI of the mutual information threshold , when the mutual information of consecutive measurement points is greater than the threshold NMI threshold , then the sensor is determined to be a redundant sensor.

9. An electronic device, characterized in that, It includes a processor and a memory. The memory stores program codes. When the program codes are executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 8.

10. A storage medium, characterized in that, Stores a computer program or instruction. When the computer program or instruction runs on a computer, it executes the steps of the method according to any one of claims 1 to 8.

Images