A method for monitoring three-dimensional state of line galloping in signal unstable area

Abstract

The application discloses a kind of line galloping three-dimensional state monitoring method for signal unstable area, it is related to information technology field, including through real-time data acquisition and VMD decomposition, the IMF component prediction value of the LSTM model of construction, the first matrix is formed by super-long component sliding value extraction.Form second matrix after invalid singular value is removed by SVD decomposition reconstruction and 0 is supplemented, average reconstruction IMF, calculate grid correction number prediction value.Combining the outline coordinates of line galloping monitoring point and the prediction value of grid correction number, generate virtual observation value, construct short baseline double difference equation, real-time Kalman filtering, obtain line galloping monitoring point coordinate time series;The method described in the application obtains the grid correction number prediction value by using VMD decomposition combined with LSTM model, weighted analysis to each IMF data principal component, improves line galloping monitoring precision;A large amount of data is processed through CORS site, and monitoring point only needs a small amount of correction number communication, effectively reduces data transmission burden and cost.

Description

Technical Field

[0001] This invention relates to the field of information technology, specifically to a method for three-dimensional monitoring of line galloping in areas with unstable signals. Background Technology

[0002] As living standards continue to improve, people's demand for electricity is also increasing. Ensuring that power transmission lines can continuously and normally supply electricity to people has become crucial. In the field of power transmission, the monitoring of line galloping plays a pivotal role in ensuring the stability and safety of power transmission lines. Line galloping refers to the vibration or swaying of power transmission lines, which may be caused by wind, rain, snow or other environmental factors. Such galloping may cause lines to come into contact with each other or collide with other structures, thereby causing power outages or power system failures.

[0003] Traditional methods for monitoring line galloping often rely on ground-based observation equipment or simple satellite-based positioning technology. While these methods can provide some monitoring capabilities, they suffer from low accuracy, slow data updates, and poor adaptability to environmental factors. In particular, under severe weather conditions, traditional methods often fail to accurately capture the three-dimensional dynamic information of the line, thus affecting the accuracy and real-time performance of the monitoring.

[0004] In existing technologies, line galloping can be monitored using network RTK. However, the high accuracy of network RTK depends on the timely broadcasting of grid corrections. If the broadcasting is not timely enough, the positioning accuracy will decrease over time due to the untimely updating of grid corrections. This poses a huge challenge for line galloping monitoring, which is usually located in areas with unstable signals. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by this invention is that traditional line galloping monitoring methods suffer from low accuracy, slow data updates, poor adaptability to environmental factors, and the problem of how to promptly broadcast grid correction data to improve the accuracy of line galloping monitoring in areas with unstable signals.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a method for three-dimensional monitoring of line galloping status in areas with unstable signals, comprising: real-time acquisition of grid correction time series data; calculation of predicted values ​​of intrinsic mode components of grid correction based on VMD decomposition; creation of a first matrix by setting a sliding window; reconstruction of the first matrix by SVD decomposition; construction of a second matrix after padding the matrix with zeros; reconstruction of intrinsic modes by taking the average value; calculation of predicted grid correction values; generation of virtual observation values ​​based on the approximate coordinates of line galloping monitoring points and predicted grid correction values; construction of a short baseline double-difference equation between the virtual observation values ​​and actual observation values; and obtaining the time series coordinates of the line galloping monitoring points through real-time Kalman filtering.

[0008] As a preferred embodiment of the three-dimensional status monitoring method for line galloping in areas with unstable signals described in this invention, wherein: the setting of the sliding window to create the first matrix includes real-time acquisition of grid correction time series data published by the line CORS station, and the data is decomposed by VMD to obtain the eigenmode components of the grid correction, {X(0)} i X(1) i ,...,X(N) i The intrinsic mode components are used to construct an LSTM model to obtain the predicted values ​​pd of the intrinsic mode components. i ;

[0009] When the length N of the intrinsic mode component is greater than the first set threshold, let there be a sliding window L. If the length of the sliding window L is less than half of N, then the sliding window starts from the beginning of the intrinsic mode component X(0). i Slide to retrieve values; obtain window data {X(t)} by sliding to retrieve values. i X(t+1) i ,...,X(t+L) i The sliding step size is 1;

[0010] The intrinsic modal components {X(t)} of the window data i X(t+1) i ,...,X(t+L) i Create the corresponding first matrix F i The first matrix F i The creation method is to use the window data {X(t)} corresponding to the intrinsic mode i. i X(t+1) i ,...,X(t+L) i The rows of the matrix form the first matrix F. i .

[0011] As a preferred embodiment of the three-dimensional status monitoring method for line galloping in areas with unstable signals described in this invention, wherein: the construction of the second matrix after padding the matrix with zeros includes modifying the first matrix F i Perform SVD decomposition, represented as:

[0012]

[0013] Among them, U i and These are the left and right singular vectors of the i-th component, respectively, Σ i Let represent the singular value matrix of the i-th component;

[0014] Σ i Singularity ρ i Sort and remove the singular value ρ that has the smallest a% among all singular values. i and the corresponding left singular vector U i and right singular vectors

[0015] Using the remaining singular values ​​ρ i * and the corresponding left singular vector U i * and right singular vectors The first matrix of the reconstructed component, and the first matrix of the reconstructed component, are represented as follows:

[0016]

[0017] Among them, F i * U is the first matrix of the reconstructed components. i * For the remaining singular values ​​ρ i * The corresponding left singular vector, For the remaining singular values ​​ρ i * The corresponding right singular vector;

[0018] Take each row of the first matrix of reconstructed intrinsic modes sequentially from top to bottom, and pad all rows with leading and trailing zeros. The padding rule is to pad the first element of the m-th row with m-1 zeros, and the L-th element of the m-th row with NL zeros. Arrange all the zero-padded rows in order to form the second matrix H. i * OK.

[0019] As a preferred embodiment of the three-dimensional monitoring method for line galloping in areas with unstable signals described in this invention, wherein: the calculation of the grid correction prediction value includes calculating the second matrix H i* The intrinsic modes are reconstructed by averaging each column.

[0020] The Euclidean distance between the reconstructed and unreconstructed intrinsic modes is calculated and expressed as:

[0021]

[0022] Among them, R i Let be the Euclidean distance of the i-th intrinsic mode component. Let X(t) be the i-th reconstructed intrinsic mode component. i Let be the intrinsic mode component before the i-th reconstruction, and t be the time point, where t≤N;

[0023] The weights of the intrinsic modal components are calculated as follows:

[0024]

[0025] Where, ρ i The weights are those of the i-th intrinsic mode component;

[0026] The predicted value of the grid correction is calculated and expressed as follows:

[0027]

[0028] Where y is the predicted value of the grid correction, and M is the number of intrinsic modes after VMD decomposition.

[0029] As a preferred embodiment of the three-dimensional monitoring method for line galloping in areas with unstable signals described in this invention, the predicted value based on the approximate coordinates and grid correction of the line galloping monitoring point includes interpolating the tropospheric delay, ionospheric delay, and comprehensive error prediction value at the line galloping monitoring point based on the approximate coordinates and grid correction prediction value y of the line galloping monitoring point.

[0030] The predicted ionospheric delay at the line galloping monitoring point is calculated. Spatial structure and spatial variability analyses are performed at the central ionospheric height. A preset height is used as the ionospheric error correction reference surface. Non-differential ionospheric delay error interpolation is performed at the line galloping monitoring point on this reference surface to obtain the ionospheric correction number. The ionospheric delay is then calculated based on this correction number. The non-differential ionospheric delay error interpolation includes obtaining the Kriging equations for interpolating the total electron content of the ionosphere under linear, unbiased, and optimal estimation conditions, expressed as:

[0031]

[0032] Where μ is the Lagrange multiplier factor, dij d represents the distance between the CORS site and the line galloping monitoring point. i0 λ is the distance between the interpolation point and the observed value. j These are weighting coefficients;

[0033] The Kriging equations are solved to obtain weighting coefficients. Based on the weighting coefficients, the estimated total electron content of the ionosphere is calculated. The predicted ionospheric delay at the line galloping monitoring point is then calculated using the estimated total electron content of the ionosphere.

[0034] The predicted tropospheric delay at the line galloping monitoring points is calculated by dividing the tropospheric correlation error into dry delay and wet delay components. The dry delay component is corrected using the GPT2w or ITG model. The wet delay component is calculated using the Kriging interpolation algorithm based on height correction, based on the tropospheric wet delay results obtained from each CORS station. The tropospheric correction number at the line galloping monitoring point is then calculated based on the tropospheric correction number at the line galloping monitoring point. The Kriging interpolation algorithm based on height correction specifically involves averaging the elevations of each CORS station to obtain the elevation datum for the tropospheric wet delay correction in the interpolation area, and then calculating the tropospheric wet delay elevation difference correction model. The tropospheric wet delay correction values ​​from each CORS station to the tropospheric error correction reference surface are obtained. Spatial structure and variability analysis of the tropospheric wet delay at each CORS station are performed on the elevation reference surface. Based on the approximate coordinates of the line galloping monitoring points and the results of the spatial structure and variability analysis of the tropospheric wet delay, the Kriging algorithm is used to interpolate the tropospheric wet delay correction values ​​from the tropospheric wet delay correction reference surface. Based on the approximate coordinates and elevation of the line galloping monitoring points, the elevation difference from the line galloping monitoring points to the tropospheric error correction reference surface is calculated. Based on the tropospheric wet delay elevation difference correction model, the predicted value of the tropospheric wet delay at the approximate coordinates and elevation of the line galloping monitoring points is calculated.

[0035] The comprehensive error prediction value at the line galloping monitoring point is calculated, and the comprehensive error prediction value is obtained by interpolation using the inverse distance weighting method.

[0036] As a preferred embodiment of the three-dimensional monitoring method for line galloping in areas with unstable signals described in this invention, the virtual observation values ​​include virtual observation values ​​of line galloping monitoring points generated based on the tropospheric delay, ionospheric delay, and comprehensive error prediction values ​​of the line galloping monitoring points.

[0037] The method for calculating the virtual observations of the line galloping monitoring point is as follows: using the coordinates and satellite ephemeris of the line galloping monitoring point and CORS station, the observations of CORS station are reduced to the line galloping monitoring point. Then, taking into account the relative difference between the predicted values ​​of tropospheric delay and ionospheric delay, the observations of the line galloping monitoring point are obtained.

[0038] The observations from CORS stations are attributed to the line galloping monitoring points and expressed as follows:

[0039] Φ' V =Φ A -ρ A +ρ V

[0040] P' V =P A -ρ A +ρ V

[0041] Where, Φ' V P' represents the carrier wave observation value at the line galloping monitoring point. V Φ represents the pseudorange observation value of the line galloping monitoring point. A For carrier observations at CORS sites, P A ρ represents the pseudorange observations at the CORS station. A ρ represents the geometric distance from the CORS station to the satellite. V The geometric distance from the line galloping monitoring point to the satellite;

[0042] Calculate the difference after double difference of tropospheric delay error at CORS sites Represented as:

[0043]

[0044] Calculate the difference after double difference of ionospheric delay error at CORS site Represented as:

[0045]

[0046] Where f1 and f2 are the frequencies of the L1 and L2 observations, respectively. and These are L1 and L2 double-difference carrier observations, respectively. and Let L1 and L2 be the double-difference integer ambiguities, and ρ be the geometric distance between the satellite and the receiver;

[0047] The difference between the two differences of the tropospheric delay error at the line galloping monitoring point was calculated using a spatial interpolation algorithm. The difference between the double difference and the ionospheric delay error Calculate the observed values ​​at the line galloping monitoring points.

[0048] As a preferred embodiment of the three-dimensional status monitoring method for line galloping in areas with unstable signals described in this invention, the method for constructing the short baseline double-difference equation includes constructing the short baseline double-difference observation equation by combining the virtual observation values ​​of the line galloping monitoring points and the actual observation values ​​of the line galloping monitoring points. The observation values ​​of the line galloping monitoring points include the carrier observation values ​​and pseudorange observation values ​​of the line galloping monitoring points.

[0049] Calculate the carrier observation value Φ at the line galloping monitoring point V , is represented as:

[0050]

[0051] in, This is the correction value for the double-difference error of the carrier observations at the line galloping monitoring point. Including the double difference error corrections for carrier observations corresponding to frequency points 1 and 2 respectively. and Represented as:

[0052]

[0053] Calculate the pseudorange observation value P at the line galloping monitoring point V , is represented as:

[0054]

[0055] in, This represents the pseudorange double-difference error correction value at the line galloping monitoring point. This includes the pseudorange double-difference error corrections for the line galloping monitoring points corresponding to frequency points 1 and 2, respectively. and Represented as:

[0056]

[0057] in, and These are the positive pseudorange double-difference errors of the line galloping monitoring points at frequencies 1 and 2, respectively. This is the difference after double-difference of ionospheric delay error.

[0058] Another objective of this invention is to provide a three-dimensional monitoring system for line galloping in areas with unstable signals. This system can analyze the influence of the principal components of each intrinsic mode data and assign corresponding weights, thereby solving the problem of low accuracy in line galloping monitoring caused by untimely updates due to signal instability in existing technologies.

[0059] As a preferred embodiment of the three-dimensional status monitoring system for line galloping in areas with unstable signals described in this invention, the system includes: a VMD decomposition module, a predicted value calculation module, and a double-difference observation module; the VMD decomposition module is used to collect grid correction time series data in real time, calculate the predicted values ​​of the intrinsic mode components of the grid correction based on VMD decomposition, and create a first matrix by setting a sliding window; the predicted value calculation module is used to perform SVD decomposition on the first matrix, reconstruct the first matrix, fill the matrix with 0s to form a second matrix, take the average to reconstruct the intrinsic modes, and calculate the predicted values ​​of the grid correction; the double-difference observation module is used to generate virtual observation values ​​based on the approximate coordinates of the line galloping monitoring points and the predicted values ​​of the grid correction, construct a short baseline double-difference equation with the virtual observation values ​​and the actual observation values, and obtain the coordinate time series of the line galloping monitoring points through real-time Kalman filtering.

[0060] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement a method for three-dimensional condition monitoring of line galloping in areas with unstable signals.

[0061] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for three-dimensional monitoring of line galloping in areas with unstable signals.

[0062] The beneficial effects of this invention are as follows: The three-dimensional monitoring method for line galloping in areas with unstable signals provided by this invention analyzes the influence of principal components of each intrinsic mode data by using VMD decomposition and constructing an LSTM model and assigning corresponding weights. The weighted summation yields the predicted value of the grid correction, thereby effectively solving the problem of decreased line galloping monitoring accuracy caused by insufficient signal stability during the calculation of existing network RTK. In the data transmission and reception process, a large amount of data is transmitted through CORS stations. The line galloping monitoring point only needs to transmit and receive a small number of corrections to monitor the line galloping situation, which not only greatly reduces the data transmission and reception pressure of a single line galloping monitoring point, but also greatly reduces the cost incurred by the monitoring point due to data transmission and reception. Attached Figure Description

[0063] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 The first embodiment of the present invention provides an overall flowchart of a three-dimensional status monitoring method for line galloping in areas with unstable signals.

[0065] Figure 2 The following is an overall flowchart of a three-dimensional status monitoring system for line galloping in areas with unstable signals, provided as a third embodiment of the present invention. Detailed Implementation

[0066] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0067] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for three-dimensional monitoring of line galloping status in areas with unstable signals is provided, comprising:

[0068] S1: Real-time acquisition of grid correction time series data, calculation of the predicted values ​​of the intrinsic mode components of the grid correction based on VMD decomposition, and creation of the first matrix by setting a sliding window.

[0069] Furthermore, the sliding window is set to create a first matrix that includes real-time acquisition of grid correction time series data published by CORS stations. The data is decomposed using VMD to obtain the eigenmode components of the grid correction, {X(0)}. i X(1) i ,...,X(N) i The intrinsic mode components are used to construct an LSTM model to obtain the predicted values ​​pd of the intrinsic mode components. i When the length N of the intrinsic mode component is greater than the first set threshold, let the sliding window L be less than half of N, then the sliding window starts from the beginning of the intrinsic mode component X(0). i Slide to retrieve values; obtain window data {X(t)} by sliding to retrieve values. i X(t+1) i ,...,X(t+L) i}, with a sliding step size of 1; for the intrinsic mode components {X(t) of the window data i X(t+1) i ,...,X(t+L) i Create the corresponding first matrix F i The first matrix F i The creation method is to use the window data X(t) corresponding to the intrinsic mode i. i X(t+1) i ,...,X(t+L) iThe rows of the matrix form the first matrix F. i .

[0070] It should be noted that VMD decomposition can decompose complex grid correction time series data into multiple intrinsic mode components. Each intrinsic mode component contains components of different frequencies and amplitudes in the original signal. This decomposition can effectively separate noise and interference components in the signal from the useful signal, thereby improving the accuracy of subsequent processing and prediction. The LSTM model can learn the long-term dependencies in the grid correction time series data and perform time series prediction. By performing LSTM model prediction on the intrinsic mode components, the grid correction prediction value at future time can be obtained, thereby realizing real-time monitoring of line galloping.

[0071] S2: Perform SVD decomposition on the first matrix, reconstruct the first matrix, fill the matrix with 0s to form the second matrix, take the average to reconstruct the eigenmodes, and calculate the predicted value of the grid correction.

[0072] Furthermore, constructing a second matrix after padding the first matrix F with zeros involves modifying the first matrix F... i Perform SVD decomposition, represented as:

[0073]

[0074] Among them, U i and These are the left and right singular vectors of the i-th component, respectively, Σ i Let Σ represent the singular value matrix of the i-th component; i Singularity ρ i Sort and remove the singular value ρ that has the smallest a% among all singular values. i and the corresponding left singular vector U i and right singular vectors Using the remaining singular values ​​ρ i * and the corresponding left singular vector U i * and right singular vectors The first matrix of the reconstructed component, and the first matrix of the reconstructed component, are represented as follows:

[0075]

[0076] Among them, F i * U is the first matrix of the reconstructed components. i * For the remaining singular values ​​ρ i * The corresponding left singular vector, For the remaining singular values ​​ρ i* The corresponding right singular vector; take each row of the first matrix of the reconstructed eigenmodes from top to bottom, and pad all the rows with 0 at the beginning and end. The padding rule is to pad the first element of the m-th row with m-1 zeros, and the L-th element of the m-th row with NL zeros. Arrange all the padded rows in order to form the second matrix H. i * OK.

[0077] It should be noted that calculating the predicted grid correction includes adjusting the second matrix H. i * The intrinsic modes are reconstructed by averaging each column. The Euclidean distance between the reconstructed and unreconstructed intrinsic modes is calculated and expressed as:

[0078]

[0079] Among them, R i Let be the Euclidean distance of the i-th intrinsic mode component. Let X(t) be the i-th reconstructed intrinsic mode component. i Let be the intrinsic mode component before the i-th reconstruction, and t be the time point, t≤N; calculate the weights of the intrinsic mode components, expressed as:

[0080]

[0081] Where, ρ i Let be the weight of the i-th intrinsic mode component; calculate the grid correction prediction value, expressed as:

[0082]

[0083] Where y is the predicted value of the grid correction, and M is the number of intrinsic modes after VMD decomposition.

[0084] It should also be noted that singular values ​​represent the energy distribution of a matrix. Smaller singular values ​​usually correspond to noise and redundant information. By removing the smallest 'a%' singular value, this noise and redundant information can be effectively removed, thereby improving the accuracy of data processing. Matrix padding with zeros is necessary because the dimensions of the reconstructed first matrix will change after removing singular values. Zero padding is required to maintain the original dimensions of the matrix. This is very important for subsequent data processing and model building, as models usually require input data with fixed dimensions.

[0085] S3: Based on the approximate coordinates of the line galloping monitoring points and the predicted grid correction values, virtual observations are generated. The virtual observations and actual observations are used to construct a short baseline double-difference equation, and the time series of line galloping monitoring point coordinates is obtained through real-time Kalman filtering.

[0086] Furthermore, based on the approximate coordinates and grid correction values ​​of the line galloping monitoring points, the predicted values ​​include interpolating the tropospheric delay, ionospheric delay, and comprehensive error prediction values ​​at the monitoring points using the approximate coordinates and grid correction values ​​y. The predicted ionospheric delay value at the monitoring points is calculated, and spatial structure and spatial variability analyses are performed at the central ionospheric height. A preset height is used as the ionospheric error correction reference surface. On this reference surface, the non-differential ionospheric delay error is interpolated to obtain the ionospheric correction value at the monitoring points. The ionospheric delay value is then calculated based on this correction. The non-differential ionospheric delay error interpolation includes obtaining the Kriging equations for interpolating the total electron content of the ionosphere under linear, unbiased, and optimal estimation conditions, expressed as:

[0087]

[0088] Where μ is the Lagrange multiplier factor, d ij d represents the distance between the CORS site and the line galloping monitoring point. i0 λ is the distance between the interpolation point and the observed value. jThe weighting coefficients are used to solve the Kriging equations. Based on these weighting coefficients, an estimate of the total ionospheric electron content is calculated. The predicted ionospheric delay at the line galloping monitoring point is then calculated from this estimate. The predicted tropospheric delay at the monitoring point is calculated, dividing the tropospheric correlation error into dry and wet delay components. The dry delay is corrected using the GPT2w or ITG model. The wet delay is calculated using the Kriging interpolation algorithm based on height correction, based on the tropospheric wet delay results from each CORS station. The tropospheric correction number at the monitoring point is then calculated, and the tropospheric delay is calculated based on this correction number. Specifically, the Kriging interpolation algorithm based on height correction involves averaging the elevations of each CORS station to obtain the tropospheric wet delay correction for the interpolation region. Based on the elevation datum, the tropospheric wet delay correction from each CORS station to the tropospheric error correction datum is calculated using the tropospheric wet delay elevation difference correction model. Spatial structural and variability analyses of the tropospheric wet delay at each CORS station are performed on the elevation datum. Based on the approximate coordinates of the line galloping monitoring points and the results of the spatial structural and variability analyses of the tropospheric wet delay, the Kriging algorithm is used to interpolate the tropospheric wet delay correction from the tropospheric wet delay datum. The elevation difference from the line galloping monitoring points to the tropospheric error correction datum is calculated based on the approximate coordinates of the line galloping monitoring points. The predicted tropospheric wet delay at the approximate coordinates of the line galloping monitoring points is calculated using the tropospheric wet delay elevation difference correction model. Finally, the predicted comprehensive error at the line galloping monitoring points is calculated and interpolated using the inverse distance weighting method.

[0089] It should be noted that the virtual observations include those generated based on the tropospheric delay, ionospheric delay, and comprehensive error prediction values ​​of the line galloping monitoring points. The calculation method for the virtual observations of the line galloping monitoring points is as follows: using the coordinates and satellite ephemeris of the line galloping monitoring points and CORS stations, the observations from the CORS stations are reduced to those of the line galloping monitoring points. Then, considering the inter-station relative differences between the predicted tropospheric delay and ionospheric delay values, the observations of the line galloping monitoring points are obtained. The reduction of the observations from the CORS stations to the line galloping monitoring points is expressed as follows:

[0090] Φ' V =Φ A -ρ A +ρ V

[0091] P' V =P A -ρ A +ρ V

[0092] Where, Φ' V P' represents the carrier wave observation value at the line galloping monitoring point. V Φ represents the pseudorange observation value of the line galloping monitoring point. A For carrier observations at CORS sites, P A ρ represents the pseudorange observations at the CORS station. A ρ represents the geometric distance from the CORS station to the satellite. V The geometric distance from the line galloping monitoring point to the satellite is given; the difference between the double differences of the tropospheric delay error at the CORS station is calculated. Represented as:

[0093]

[0094] Calculate the difference after double difference of ionospheric delay error at CORS site Represented as:

[0095]

[0096] Where f1 and f2 are the frequencies of the L1 and L2 observations, respectively. and These are L1 and L2 double-difference carrier observations, respectively. and Let L1 and L2 be the double-difference integer ambiguities, and ρ be the geometric distance between the satellite and the receiver. The difference in tropospheric delay error at the line galloping monitoring point after double-difference is calculated using a spatial interpolation algorithm. The difference between the double difference and the ionospheric delay error Calculate the observed values ​​at the line galloping monitoring points.

[0097] It should also be noted that the construction of the short baseline double-difference equation involves constructing a short baseline double-difference observation equation using the virtual and actual observations of the line galloping monitoring points. The observations of the line galloping monitoring points include both carrier observations and pseudorange observations. The carrier observation value Φ of the line galloping monitoring points is then calculated. V , is represented as:

[0098]

[0099] in, This is the correction value for the double-difference error of the carrier observations at the line galloping monitoring point. Including the double difference error corrections for carrier observations corresponding to frequency points 1 and 2 respectively. and Represented as:

[0100]

[0101] Calculate the pseudorange observation value P at the line galloping monitoring point V , is represented as:

[0102]

[0103] in, This represents the pseudorange double-difference error correction value at the line galloping monitoring point. This includes the pseudorange double-difference error corrections for the line galloping monitoring points corresponding to frequency points 1 and 2, respectively. and Represented as:

[0104]

[0105] in, and These are the positive pseudorange double-difference errors of the line galloping monitoring points at frequencies 1 and 2, respectively. This is the difference after double-difference of ionospheric delay error.

[0106] It should also be noted that the short baseline double-difference equation constructed from virtual and actual observations can effectively eliminate or reduce common errors in the observation data, such as satellite orbit errors and clock errors. At the same time, by using a Kalman filter model to filter the short baseline double-difference observation equation in real time, the accuracy and real-time performance of line galloping monitoring can be further improved. The Kalman filter model can effectively integrate observation data and prior information to obtain a more accurate time series of coordinates of line galloping monitoring points.

[0107] Example 2 is an embodiment of the present invention, which provides a three-dimensional monitoring method for line galloping in areas with unstable signals. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0108] First, the purpose of the experiment is to verify the accuracy and effectiveness of the present invention in three-dimensional monitoring of line galloping in areas with unstable signals, and to compare it with existing technologies.

[0109] Experimental Procedure: Grid correction time-series data were collected in real-time from CORS sites at 1-second intervals for a total of 1000 seconds. The VMD algorithm was used to decompose the data, yielding 5 intrinsic mode components (IMFs). An LSTM model was built using the TensorFlow framework, with the 5 IMFs as input and their predictions as output. Model training parameters were set as follows: 50 hidden layer neurons, a learning rate of 0.001, a batch size of 64, and 100 training epochs. The threshold N was set to 100, and the sliding window length L was set to 50. Sliding window values ​​were applied to IMFs with a length greater than 100 to create the first matrix. The first matrix was then decomposed using SVD, removing the smallest 10% of singular values ​​and their corresponding... The left and right singular vectors are used to reconstruct the first matrix, which is then padded with zeros to form the second matrix. The average of each column of the second matrix is ​​taken to reconstruct the eigenmodes, and the grid correction prediction value is calculated. Error prediction is performed using the calculated grid correction prediction value to predict the ionospheric delay, tropospheric delay, and overall error. Using the coordinates and satellite ephemeris of the line galloping monitoring points and CORS stations, the observation values ​​from the CORS stations are reduced to the line galloping monitoring points. Considering the inter-station relative difference between the predicted tropospheric delay and the predicted ionospheric delay, the observation values ​​of the line galloping monitoring points are obtained. The virtual and actual observation values ​​of the line galloping monitoring points are used to construct a short-baseline double-difference equation, and a Kalman filter model is used for real-time Kalman filtering to obtain the time series of the line galloping monitoring point coordinates.

[0110] The accuracy of the time series of coordinates of the line galloping monitoring points obtained by the method of the present invention is compared with that of the time series of coordinates obtained by the traditional Kalman filtering method, and the positioning error is calculated as shown in Table 1 below.

[0111] Table 1 Comparison of data between the present invention and prior art.

[0112]

[0113] As can be seen from the table data, our invention outperforms existing technologies in terms of ionospheric delay prediction error, tropospheric delay prediction error, comprehensive error prediction error, and positioning error. This is mainly because our invention uses VMD decomposition and LSTM model, which can effectively extract and predict the main components in the grid correction, thereby improving the accuracy of error prediction. In addition, our invention further eliminates noise and redundant information by using sliding window value selection and SVD decomposition, thereby improving the efficiency and accuracy of data processing.

[0114] Example 3, referring to Figure 2 As an embodiment of the present invention, a three-dimensional monitoring system for line galloping in areas with unstable signals is provided, including a VMD decomposition module, a predicted value calculation module, and a double-difference observation module.

[0115] The VMD decomposition module is used to collect grid correction time series data in real time. Based on VMD decomposition, it calculates the predicted values ​​of the intrinsic mode components of the grid correction and sets a sliding window to create a first matrix. The prediction value calculation module is used to perform SVD decomposition on the first matrix, reconstruct the first matrix, fill the matrix with zeros to form a second matrix, take the average to reconstruct the intrinsic modes, and calculate the predicted value of the grid correction. The double-difference observation module is used to generate virtual observation values ​​based on the approximate coordinates of the line galloping monitoring points and the predicted values ​​of the grid correction. The virtual observation values ​​and the actual observation values ​​form a short baseline double-difference equation, and the coordinate time series of the line galloping monitoring points is obtained through real-time Kalman filtering.

[0116] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0117] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0118] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0119] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc. It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for three-dimensional monitoring of line galloping in areas with unstable signals, characterized in that, include: Real-time acquisition of grid correction time series data; calculation of the predicted values ​​of the intrinsic mode components of grid correction based on VMD decomposition; and creation of the first matrix by setting a sliding window. The first matrix is ​​decomposed by SVD, the first matrix is ​​reconstructed, the matrix is ​​padded with zeros to form the second matrix, the average value is taken to reconstruct the eigenmodes, and the grid correction prediction value is calculated. Based on the approximate coordinates of the line galloping monitoring points and the predicted values ​​of the grid correction, virtual observations are generated. The virtual observations and the actual observations are used to construct a short baseline double difference equation, and the time series of the line galloping monitoring point coordinates is obtained through real-time Kalman filtering. The first matrix created by setting a sliding window includes real-time acquisition of grid correction time series data published by CORS stations. The data is decomposed using VMD to obtain the eigenmode components of the grid corrections. An LSTM model is constructed using the intrinsic mode components to obtain the predicted values ​​of the intrinsic mode components. ; When the length N of the intrinsic mode component is greater than a first set threshold, let there be a sliding window L. If the length of the sliding window L is less than half of N, then the sliding window starts from the beginning of the intrinsic mode component. Slide to retrieve values, retrieve window data by sliding. The sliding step size is 1; intrinsic mode components of window data Create the corresponding first matrix The first matrix The creation method is to use the window data corresponding to intrinsic mode i. The rows of the matrix form the first matrix. .

2. The method for three-dimensional monitoring of line galloping in areas with unstable signals as described in claim 1, characterized in that: The process of padded with zeros to construct the second matrix includes modifying the first matrix... Perform SVD decomposition, represented as: in, and Let be the left singular vector and the right singular vector of the i-th component, respectively. Let represent the singular value matrix of the i-th component; Will Singularity Sort and remove the singular value with the smallest a% among all singular values. and the corresponding left singular vector and right singular vector ; Utilize the remaining singular values and the corresponding left singular vector and right singular vector The first matrix of the reconstructed component, and the first matrix of the reconstructed component, are represented as follows: in, This is the first matrix of the reconstructed components. For the remaining singular values The corresponding left singular vector, For the remaining singular values The corresponding right singular vector; Take each row of the reconstructed intrinsic mode matrix from top to bottom, and pad all rows with zeros at the beginning and end. The padding rule is to pad the first element of the m-th row with m-1 zeros, and the L-th element of the m-th row with NL zeros. Arrange all the zero-padded rows in order to form the second matrix. OK.

3. The method for three-dimensional monitoring of line galloping in areas with unstable signals as described in claim 2, characterized in that: The calculation of the grid correction prediction value includes the second matrix. The intrinsic modes are reconstructed by averaging each column. ; The Euclidean distance between the reconstructed and unreconstructed intrinsic modes is calculated and expressed as: in, Let be the Euclidean distance of the i-th intrinsic mode component. For the i-th reconstructed intrinsic mode component, Let t be the intrinsic mode component before the i-th reconstruction, and t be the time point. ; The weights of the intrinsic modal components are calculated as follows: in, The weights are those of the i-th intrinsic mode component; The predicted value of the grid correction is calculated and expressed as follows: Where y is the predicted value of the grid correction, and M is the number of intrinsic modes after VMD decomposition.

4. The method for three-dimensional monitoring of line galloping in areas with unstable signals as described in claim 3, characterized in that: The predicted value based on the approximate coordinates and grid correction of the line galloping monitoring point includes interpolating the tropospheric delay, ionospheric delay, and comprehensive error prediction value at the line galloping monitoring point based on the approximate coordinates and grid correction prediction value y of the line galloping monitoring point. The predicted ionospheric delay at the line galloping monitoring point is calculated. Spatial structure and spatial variability analyses are performed at the central ionospheric height. A preset height is used as the ionospheric error correction reference surface. Non-differential ionospheric delay error interpolation is performed at the line galloping monitoring point on this reference surface to obtain the ionospheric correction number. The ionospheric delay is then calculated based on this correction number. The non-differential ionospheric delay error interpolation includes obtaining the Kriging equations for interpolating the total electron content of the ionosphere under linear, unbiased, and optimal estimation conditions, expressed as: in, For the Lagrange multiplier factor, The distance between the CORS site and the line galloping monitoring point. The distance between the interpolation point and the observed value. These are weighting coefficients; The Kriging equations are solved to obtain weighting coefficients. Based on the weighting coefficients, the estimated total electron content of the ionosphere is calculated. The predicted ionospheric delay at the line galloping monitoring point is then calculated using the estimated total electron content of the ionosphere. The predicted tropospheric delay at the line galloping monitoring points is calculated by dividing the tropospheric correlation error into dry delay and wet delay components. The dry delay component is corrected using the GPT2w or ITG model. The wet delay component is calculated using the Kriging interpolation algorithm based on height correction, based on the tropospheric wet delay results obtained from each CORS station. The tropospheric correction number at the line galloping monitoring point is then calculated based on the tropospheric correction number at the line galloping monitoring point. The Kriging interpolation algorithm based on height correction specifically involves averaging the elevations of each CORS station to obtain the elevation datum for the tropospheric wet delay correction in the interpolation area, and then calculating the tropospheric wet delay elevation difference correction model. The tropospheric wet delay correction values ​​from each CORS station to the tropospheric error correction reference surface are obtained. Spatial structure and variability analysis of the tropospheric wet delay at each CORS station are performed on the elevation reference surface. Based on the approximate coordinates of the line galloping monitoring points and the results of the spatial structure and variability analysis of the tropospheric wet delay, the Kriging algorithm is used to interpolate the tropospheric wet delay correction values ​​from the tropospheric wet delay correction reference surface. Based on the approximate coordinates and elevation of the line galloping monitoring points, the elevation difference from the line galloping monitoring points to the tropospheric error correction reference surface is calculated. Based on the tropospheric wet delay elevation difference correction model, the predicted value of the tropospheric wet delay at the approximate coordinates and elevation of the line galloping monitoring points is calculated. The comprehensive error prediction value at the line galloping monitoring point is calculated, and the comprehensive error prediction value is obtained by interpolation using the inverse distance weighting method.

5. The method for three-dimensional monitoring of line galloping in areas with unstable signals as described in claim 4, characterized in that: The virtual observations include virtual observations of the line galloping monitoring points generated based on the tropospheric delay, ionospheric delay, and comprehensive error prediction values ​​of the monitoring points. The method for calculating the virtual observations of the line galloping monitoring point is as follows: using the coordinates and satellite ephemeris of the line galloping monitoring point and CORS station, the observations of CORS station are reduced to the line galloping monitoring point. Then, taking into account the relative difference between the predicted values ​​of tropospheric delay and ionospheric delay, the observations of the line galloping monitoring point are obtained. The observations from CORS stations are attributed to the line galloping monitoring points and expressed as follows: in, For the carrier wave observations at the line galloping monitoring point, The pseudorange observations are from the line galloping monitoring points. For carrier observations at CORS sites, These are pseudorange observations from CORS stations. The geometric distance from the CORS site to the satellite. The geometric distance from the line galloping monitoring point to the satellite; Calculate the difference after double difference of tropospheric delay error at CORS sites , is represented as: Calculate the difference after double difference of ionospheric delay error at CORS site , is represented as: in, and The frequencies of L1 and L2 observations are respectively. and These are L1 and L2 double-difference carrier observations, respectively. and For L1 and L2 double-difference integer ambiguities, The geometric distance between the satellite and the receiver; The difference between the two differences of the tropospheric delay error at the line galloping monitoring point was calculated using a spatial interpolation algorithm. The difference between the double difference and the ionospheric delay error Calculate the observed values ​​at the line galloping monitoring points.

6. The method for three-dimensional monitoring of line galloping in areas with unstable signals as described in claim 5, characterized in that: The construction of the short baseline double-difference equation includes constructing a short baseline double-difference observation equation using the virtual observation values ​​and the actual observation values ​​of the line galloping monitoring points. The observation values ​​of the line galloping monitoring points include the carrier observation values ​​and pseudorange observation values ​​of the line galloping monitoring points. Calculate carrier observations at line galloping monitoring points , is represented as: in, This is the correction value for the double-difference error of the carrier observations at the line galloping monitoring point. Including the double difference error corrections for carrier observations corresponding to frequency points 1 and 2 respectively. and , is represented as: Calculate the pseudorange observations at the line galloping monitoring points , is represented as: in, This represents the pseudorange double-difference error correction value at the line galloping monitoring point. This includes the pseudorange double-difference error corrections for the line galloping monitoring points corresponding to frequency points 1 and 2, respectively. and , is represented as: in, and These are the positive pseudorange double-difference errors of the line galloping monitoring points at frequency points 1 and 2, respectively. This is the difference after double-difference of ionospheric delay error.

7. A system employing the three-dimensional condition monitoring method for line galloping in areas with unstable signals as described in any one of claims 1 to 6, characterized in that: Includes VMD decomposition module, predicted value calculation module, and double-difference observation module; The VMD decomposition module is used to collect grid correction time series data in real time, calculate the predicted values ​​of the intrinsic mode components of the grid correction based on VMD decomposition, and set a sliding window to create the first matrix. The prediction value calculation module is used to perform SVD decomposition on the first matrix, reconstruct the first matrix, fill the matrix with 0 to form the second matrix, take the average to reconstruct the eigenmode, and calculate the grid correction prediction value. The double-difference observation module is used to generate virtual observations based on the approximate coordinates of the line galloping monitoring points and the predicted values ​​of the grid corrections. The virtual observations and the actual observations form a short-baseline double-difference equation, and the coordinate time series of the line galloping monitoring points is obtained through real-time Kalman filtering.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the three-dimensional status monitoring method for line galloping in areas with unstable signals, as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the three-dimensional status monitoring method for line galloping in areas with unstable signals, as described in any one of claims 1 to 6.

Images