Stress time sequence change point-based line icing thickness estimation method and system, and medium

By collecting and analyzing the multi-time sequence data of the distribution line, combining nuclear principal component analysis and Bayesian online change point detection, the problem of low accuracy of ice-covering monitoring of distribution line in the existing technology is solved, and higher coverage is achieved Ice thickness calculation accuracy.

CN118940020AInactive Publication Date: 2024-11-12SICHUAN ZHONGDIAN AOSTAR INFORMATION TECHNOLOGIES CO LTD

Patent Information

Application Number
CN202411418757.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-12
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing distribution line ice covering monitoring methods have low accuracy and are susceptible to factors such as smoke and light. The difference between the outer material and the line of the fiber ice covering monitoring method leads to low calculation accuracy.

Method used

The line ice-cover thickness estimation method based on stress timing change points is adopted. Multivariate timing data are collected for core principal components analysis, the initial value of ice-cover thickness is calculated, and the stress timing analysis is performed when the ice-covering conditions are met, and Bayesian online variation is used. Point detection determines the thickness of ice.

Benefits of technology

The accuracy of the calculation of the ice thickness of the distribution line is improved, the impact of abnormal data on the calculation is eliminated, and the data quality and calculation accuracy are ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940020A_ABST
    Figure CN118940020A_ABST
Patent Text Reader

Abstract

The invention discloses a line icing thickness estimation method and system based on stress time sequence change points and a medium, and the method comprises the steps: carrying out the kernel principal component analysis of collected multivariate time sequence data, and obtaining the key factors of line icing; based on the key factors of line icing, calculating an initial value of the thickness of icing, and if the initial value of the thickness of icing is greater than or equal to a set threshold value, considering that the basic condition of icing is met; and carrying out icing stress time sequence analysis, detecting whether the icing stress time sequence has a change point or not, if the time sequence change point exists, judging that the line is iced, and calculating the icing thickness of the line. According to the method, under the condition of considering the wind speed, the icing thickness of the line is calculated through the wire and the icing density, the method has the characteristic of high accuracy, and the problem of low DL icing thickness calculation accuracy in the prior art is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of power distribution line monitoring, and in particular relates to a method, system and medium for estimating line ice thickness based on stress time series change points. Background Art

[0002] Distribution lines (DL) refer to the lines that transmit electricity from substations to user power equipment. They are important infrastructure for China's social economy and are characterized by wide geographical distribution and large number. In recent years, with the continuous intensification of global climate change, the frequency of extreme low temperature weather has gradually increased. The frequency of accidents such as DL disconnection and distribution tower collapse caused by ice disasters has increased, which has brought great challenges to the stable operation of DL. Therefore, exploring the calculation of DL ice thickness and initiating ice melting has a positive effect on improving the reliability of DL power supply in ice and snow weather.

[0003] At present, many scholars have done a lot of research on DL ice warning, which is mainly divided into two categories: ice image recognition method and optical fiber ice monitoring method. Among them, the ice image recognition method uses the camera installed on the DL to collect line images, uses the image edge detection algorithm to obtain the image contour value before and after DL ice, and then performs a difference operation on the image contour value before and after DL ice to obtain the DL ice thickness. However, the ice image recognition method is easily affected by factors such as smoke and light, and the DL ice thickness measurement error is large.

[0004] In the optical fiber icing monitoring method, a simulated optical fiber for icing calculation is set up on the DL, and an optical fiber tension tilt sensor is used to measure the inclination angle and tension value at the hanging point. On this basis, the DL icing thickness is inferred by calculating the icing thickness of the optical fiber. However, the outer material, tension change value, icing rate, etc. of the optical fiber are far different from those of the DL line, and the accuracy of DL icing thickness calculation is low.

[0005] In summary, there are various DL ice monitoring methods at present, but due to technical limitations, the accuracy of DL ice thickness calculation is low. Summary of the invention

[0006] The purpose of the present invention is to provide a method, system and medium for estimating line ice thickness based on stress time series change points, aiming to solve the above-mentioned problems.

[0007] The present invention is mainly achieved through the following technical solutions: The method for estimating line ice thickness based on stress time series change points includes the following steps: Step S1: Collect multivariate time series data that affect line icing; Step S2: performing kernel principal component analysis on the multivariate time series data collected in step S1 to obtain key factors of line icing; Step S3: Based on the key factors of line icing, calculate the initial value of ice thickness; if the initial value of ice thickness is greater than or equal to the set threshold, it is considered that the basic conditions for icing are met, and enter step S4; Step S4: Perform ice stress time series analysis; when considering wind deviation, the stress value of the line F WIND for: (8) in: F BM It is the stress reference value when the line is not covered with ice and there is no wind deflection; β is the elevation difference angle of the line; i is the windage angle; Icing stress on the line over time for: (9) in: t APS It is a time series of ice cover statistics; Step S5: Detect whether there is a change point in the ice stress time series. If there is a time series change point, it is determined that the line is iced and the thickness of the ice on the line is calculated. o ICI : (12) in: P LINE is the density of line conductors; P ICE is the density of ice on the line; o LINE is the wire diameter of the line conductor; π is the ratio of pi.

[0008] In order to better implement the present invention, further, step S1 includes the following steps: Step S11: collecting multivariate time series data affecting ice coverage; Step S12: Evaluate the quality of the collected multivariate time series data. If the data quality meets the standard, proceed to step S2; otherwise, proceed to step S13; Step S13: Repair the multivariate time series data, and then proceed to step S2.

[0009] In order to better implement the present invention, further, in step S12, the consistency, completeness, accuracy, timeliness and effectiveness of the multivariate time series data are detected according to the evaluation rules, and the normal data rate d is calculated.NRM for: (1) in: d NE The amount of non-null data in the integrity check; d NULL The amount of data with empty values ​​in integrity check; n is the number of evaluation rules; is the amount of abnormal data under different evaluation rules; If the normal data rate d NRM If the value is greater than or equal to the set threshold, the data quality is considered to meet the standard; otherwise, the quantity quality is considered to be unsatisfactory.

[0010] In order to better implement the present invention, further, in step S13, the multivariate time series data is repaired based on the least squares method, and the repaired time series data d REP for: (2) in: d NVH is the original sequence of data collection; l is the slope of the original series from which the data was collected; d The original sequence of data collection is y The intercept of the axis.

[0011] In order to better implement the present invention, further, in step S2, the contribution rate is calculated or : (5) in: is the covariance matrix of multivariate time series data of different dimensions; n S is the number of eigenvectors of multivariate time series data; is the covariance matrix of multivariate time series data with different eigenvectors; The multivariate time series data are sorted from large to small according to the contribution rate to obtain the key factors of icing on the top-ranked lines.

[0012] In order to better implement the present invention, further, in step S2, the key factors for obtaining line icing include air temperature, humidity and wind speed; in step S3, the initial value of the ice thickness is: (6) in: oh It is the function for calculating the ice thickness of the line; t ATM is the current line air temperature; t O The air temperature is the freezing point; r RH is the air humidity of the line; v WIND is the wind speed of the line.

[0013] Specifically, oh ( t ATM - t O )= K ×( t ATM - t O )×(1+ α × v WIND ), in: α and K is the empirical coefficient.

[0014] In order to better implement the present invention, further, in step S4, the stress reference value of the line F BM for: (7) in: is the unit deadweight load of the line conductor; is the spacing between line towers; l LE is the length of the line.

[0015] In order to better implement the present invention, further, in step S5, a Bayesian online change point detection method is used to detect whether there is a change point in the ice stress time series.

[0016] The present invention is mainly achieved through the following technical solutions: The line ice thickness estimation system based on stress time series change point is based on the above-mentioned line ice thickness estimation method based on stress time series change point, and includes a data acquisition and processing module, a key factor analysis module, an ice stress time series analysis module and an ice thickness evaluation module; the data acquisition and processing module is used to collect multivariate time series data affecting ice coating, and perform data preprocessing, and then send it to the key factor analysis module; the key factor analysis module is used to perform kernel principal component analysis on the multivariate time series data, obtain the key factors of line ice coating, calculate the initial value of ice coating thickness, so as to determine whether the basic conditions for ice coating are met; the ice stress time series analysis module is used to calculate the ice stress time series when the basic conditions for ice coating are met, and detect the change points of the stress time series; the ice thickness evaluation module is used to calculate the ice thickness of the line.

[0017] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned line ice thickness estimation method based on stress time series change points.

[0018] The beneficial effects of the present invention are as follows: The present invention pre-processes the collected multivariate time series data such as DL tension value, wind deflection angle value, temperature, wind speed, humidity value, etc., to eliminate the influence of abnormal data on ice thickness calculation; secondly, the present invention obtains key factors through kernel principal component analysis, and then calculates ice conditions according to the key factors; when ice conditions are met, Bayesian online change point detection is performed on the DL ice stress time series to determine whether DL is iced; if iced, the DF ice thickness is calculated by the conductor and ice density while considering the wind speed, which has the characteristics of high accuracy and solves the problem of low accuracy in DL ice thickness calculation in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flow chart of the line ice thickness estimation method based on stress time series change points of the present invention; Figure 2 It is a structural schematic diagram of the ice-covering monitoring status of the distribution line; Figure 3 Schematic diagram of quality assessment for multivariate time series data; Figure 4 Schematic diagram of the wind deflection plane of the distribution line. DETAILED DESCRIPTION

[0020] Embodiment 1: Line ice thickness estimation method based on stress time series change points, such as Figure 1 As shown, the following steps are included: ①DL multivariate data collection and repair First, the line tension value and inclination angle are collected through the line tension sensor and insulator tilt sensor respectively, and the temperature, humidity and wind speed data are collected through the micro-meteorological device; secondly, the quality of the collected multivariate data is evaluated to determine whether the data quality meets the standard for DL ​​ice thickness calculation; if it meets the standard, go directly to the next step; if it does not meet the standard, repair the DL tension, inclination angle and other data that do not meet the standard, and then go to the next step.

[0021] ②Analysis of factors affecting DL icing stress Perform kernel principal component analysis (KPCA) on the environmental data of DL to obtain the types of environmental factors that affect DL icing. On this basis, analyze the DL icing conditions to obtain the environmental values ​​that form the DL icing conditions. If the environmental value is greater than or equal to the threshold, the basic icing conditions are met and proceed to the next step; otherwise, end.

[0022] ③DL ice thickness calculation First, the DL ice stress baseline value is calculated, and the DL ice stress time series analysis is performed on the DL that meets the basic environmental conditions for ice to obtain the DL stress change value; secondly, the DL ice stress time series change point value is calculated to determine whether there is a stress time series change point. If there is a stress time series change point, it is considered that there is ice, and the ice thickness of the stress time series change point DL is calculated.

[0023] Preferably, the specific steps of DL multivariate data collection and repair are as follows: (1) DL multivariate data collection: like Figure 2 As shown in the figure, in the monitoring of ice thickness of distribution lines, line tension sensors, insulator tilt sensors and micro-meteorological monitoring devices are used to collect line tension values, insulator tilt angles, and meteorological data such as temperature, humidity, and wind speed. The line tension sensor is installed on the insulator string to monitor the tension changes of the conductor. The insulator tilt sensor is fixed horizontally on the insulator, and the inclination change of the insulator relative to the horizontal plane is measured by "liquid pendulum", "solid pendulum" and "gas pendulum". The micro-meteorological monitoring device is installed on the distribution line tower, and the wind speed is measured by ultrasonic method, and the DL environmental temperature and humidity are measured by electronic temperature and humidity sensors. The data collected by the line tension sensor, insulator tilt sensor and micro-meteorological monitoring device are concentrated by using a convergence unit, and the data is transmitted to the distribution digital cloud platform through a dedicated access point (APN) for DL ​​ice monitoring.

[0024] (2) DL multivariate data quality assessment: The quality of DL multivariate data directly affects the accuracy of DL ice thickness calculation. At present, line tension sensors, insulator tilt sensors and micro-meteorological monitoring devices are easily interfered by noise during data collection and data transmission by the aggregation unit, resulting in data loss, errors and other anomalies. Therefore, it is necessary to evaluate the quality of DL multivariate data to avoid abnormal data affecting the accuracy of DL ice thickness calculation.

[0025] like Figure 3 As shown, the characteristics of the data are evaluated according to the evaluation rules, and the evaluation rules include format check, null value check, range check, timely check and logic check, and the evaluation characteristics correspond to consistency, completeness, accuracy, timeliness and effectiveness. In evaluating the quality of DL multivariate data, the format check is used to determine whether the DL multivariate data follows a unified format specification, and its consistency index is evaluated; the null value check is used to evaluate whether the DL multivariate data is missing, and its completeness index is evaluated; the range check is used to evaluate whether the data record value is within a reasonable range, and its accuracy index is evaluated; the timely check is used to measure the time interval from the generation of the DL multivariate data to the queryability, and its timeliness index is evaluated; the logical check is used to evaluate whether the DL multivariate data value meets the line icing definition requirements, and its effectiveness index is evaluated.

[0026] After evaluating the quality of DL multivariate data, the normal data rate d NRM for: (1) in: d NE DL multivariate data with non-null values; d NULL DL multivariate data with null values; n is the number of evaluation rules; is the abnormal data under different evaluation rules.

[0027] If the normal data rate d NRM If it is greater than or equal to the threshold, the data quality is considered to be up to standard, otherwise data repair is performed.

[0028] (3) DL multivariate data repair: By repairing the erroneous DL multivariate data, the error of abnormal data in DL ice thickness measurement can be reduced. Least squares (LS) is a mathematical estimation method of error. LS calculates the best function match of the erroneous data through the minimum error score, and minimizes the sum of the squares of the difference between the estimated DL multivariate result data and the actual DL multivariate data. LS has the characteristics of high calculation accuracy and fast speed. Therefore, LS is used to repair various types of erroneous data. The DL multivariate data repaired by LSd REP for: (2) in: d NVH The original sequence for DL ​​multivariate data collection; l is the original series slope collected for DL ​​multivariate data; d The original sequence collected for DL ​​multivariate data is y The intercept of the axis.

[0029] Preferably, the specific steps of analyzing the factors affecting DL icing stress are as follows: (1) Analysis of factors affecting DL icing After collecting line tension value, insulator tilt angle, temperature and humidity, and wind speed meteorological data through sensors, KPCA analysis is performed on them to obtain the key factors affecting DL icing.

[0030] KPCA is a nonlinear data processing method that nonlinearly maps the DL multivariate data in the original space to a high-dimensional feature space, and then performs principal component analysis on it to obtain the key factors affecting DL icing.

[0031] First, KPCA uses a nonlinear mapping method to standardize the DL multivariate data that affects DL icing to form a matrix and map it to a high-dimensional feature space, specifically: (3) in: , Respectively j , k The multivariate DL multivariate data after group repair; K is the kernel function; c is the KPCA mapping function; Then the DL multivariate data is calculated to obtain the standardized covariance matrix E COV .

[0032] (4) in: m is the dimension of the covariance matrix; Z The number of hours counted for DL; is DL multivariate data of different dimensions.

[0033] Finally, calculate the contribution rate of KPCA or : (5) in: is the covariance matrix of DL multivariate data of different dimensions;n S is the number of DL multivariate data feature vectors; is the covariance matrix of DL multivariate data with different eigenvectors.

[0034] The DL multivariate data are sorted from large to small according to the contribution rate to obtain the main factors of DL icing, specifically air temperature, humidity, and wind speed.

[0035] (2) Analysis of DL icing meteorological conditions It can be seen from the above that the formation of DL ice is related to meteorological conditions such as air temperature, humidity, and wind speed. When the ambient air temperature of the DL is below the freezing point, the cooling water droplets will freeze on the DL surface and form ice. Under this condition, the greater the wind speed and air humidity, the more serious the DL ice. The initial value of the thickness of the DL ice is b ICE for: (6) in: oh is the calculation function of DL ice thickness; t ATM is the current DL air temperature; t O DL air temperature for the freezing point; r RH is the air humidity of DL; v WIND is the wind speed of DL. If the initial value of ice thickness is greater than or equal to the set threshold, it is considered that the basic conditions for ice coverage are met.

[0036] Preferably, the DL ice thickness calculation specifically includes the following steps: In DL, due to the different diameters and materials of the conductors and the differences in tension sensors, the unified calculation deviation of ice thickness is large. Therefore, the DL ice thickness calculation is performed based on the time series of DL ice stress changes.

[0037] (1) Time series analysis of DL ice stress: The DL stress reference value refers to the stress value when the DL is not covered with ice and there is no wind deflection. The tension of the DL is the self-weight of the wire, which is evenly distributed along the wire suspension point and can be solved using the parabola formula. F BM for: (7) in: is the unit deadweight load of the DL conductor; is the spacing between DL towers; β is the height difference angle of DL; l LEis the length of the DL line.

[0038] like Figure 4 As shown in the figure, when the line is affected by wind, the plane formed by DL and insulators is offset from the vertical plane of the geography as a whole and forms a certain angle. The offset plane is the DL wind deviation plane, and the wind deviation angle is i , the height difference angle of DL is β After wind deflection, the DL stress value F WIND for: (8) DL ice stress changing with time for: (9) in: t APS It is a time series of ice cover statistics.

[0039] (2) DL ice change point analysis: A change point refers to a point where a significant change occurs in the DL icing time series. By detecting the change point, the change in DL tension can be effectively detected and an early warning of the icing situation can be issued.

[0040] Bayesian online changepoint detection (BOCD) is an online changepoint detection method that focuses on result prediction filtering and predicts the exact distribution of data in the generated sequence after considering the observed DL ice stress change data.

[0041] The prior probability of DL ice change point detection is p AP for: (10) in: p is the probability distribution; r It is the variable length data in BOCD; is the DL ice stress sequence at the next moment.

[0042] The posterior probability of DL ice change point detection is p LP for: (11) r The probability depends on when the change point occurs, that is, when the line application that generates DL icing changes, it can be judged that DL icing occurs.

[0043] (3) DL ice thickness analysis When DL is covered, the tension sensor will bear the weight of the overhead lines and ice on the left and right sides. The ice tension of the DL line is affected by the wind deflection angle. ϴ, The tensile force before and after ice coating is related to the ice coating thickness of DL. o ICI for: (12) in: P LINE is the density of DL line conductors; P ICE is the density of ice on DL; o LINE is the wire diameter of the DL line conductor; π is the ratio of pi. According to the above formula, the thickness of ice cover DL is calculated.

[0044] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. Line ice thickness estimation method based on stress time series change points, It is characterized in that The following steps are involved: Step S1: Collect multivariate time series data that affect line icing; Step S2: performing kernel principal component analysis on the multivariate time series data collected in step S1 to obtain key factors of line icing; Step S3: Calculate the initial value of ice thickness based on the key factors of line ice coverage; If the initial value of the ice thickness is greater than or equal to the set threshold, it is considered that the basic conditions for ice are met, and the process goes to step S4; Step S4: Perform ice stress time series analysis; when considering wind deviation, the stress value of the line F WIND for: (8) in: F BM It is the stress reference value when the line is not covered with ice and there is no wind deflection; β is the elevation difference angle of the line; θ is the windage angle; Icing stress on the line over time for: (9) in: t APS It is a time series of ice cover statistics; Step S5: Detect whether there is a change point in the ice stress time series. If there is a time series change point, it is determined that the line is iced and the thickness of the ice on the line is calculated. o ICI : (12) in: Ψ LINE is the density of line conductors; Ψ ICE is the density of ice on the line; o LINE is the wire diameter of the line conductor; π is the ratio of pi.

2. According to the method for estimating line ice thickness based on stress time series change points in claim 1, It is characterized in that The step S1 comprises the following steps: Step S11: collecting multivariate time series data affecting ice coverage; Step S12: Evaluate the quality of the collected multivariate time series data. If the data quality meets the standard, proceed to step S2; otherwise, proceed to step S13; Step S13: Repair the multivariate time series data, and then proceed to step S2.

3. According to the method for estimating line ice thickness based on stress time series change points in claim 2, It is characterized in that In step S12, the consistency, completeness, accuracy, timeliness and effectiveness of the multivariate time series data are detected according to the evaluation rules, and the normal data rate d is calculated. NRM for: (1) in: d NE The amount of non-null data in the integrity check; d NULL The amount of data with empty values ​​in integrity check; n is the number of evaluation rules; is the amount of abnormal data under different evaluation rules; If the normal data rate d NRM If the value is greater than or equal to the set threshold, the data quality is considered to meet the standard; otherwise, the quantity quality is considered to be unsatisfactory.

4. According to the method for estimating line ice thickness based on stress time series change points of claim 3, It is characterized in that In step S13, the multivariate time series data are repaired based on the least squares method. d REP for: (2) in: d NVH is the original sequence of data collection; λ is the slope of the original series from which the data was collected; δ The original sequence of data collection is y The intercept of the axis.

5. According to the method for estimating line ice thickness based on stress time series change points in claim 1, It is characterized in that In step S2, the contribution rate is calculated η : (5) in: is the covariance matrix of multivariate time series data of different dimensions; n S is the number of eigenvectors of multivariate time series data; is the covariance matrix of multivariate time series data with different eigenvectors; The multivariate time series data are sorted from large to small according to the contribution rate to obtain the key factors of icing on the top-ranked lines.

6. According to the method for estimating line ice thickness based on stress time series change points of claim 5, It is characterized in that In step S2, the key factors for obtaining line icing include air temperature, humidity and wind speed; in step S3, the initial value of the ice thickness is: (6) in: ω It is the function for calculating the ice thickness of the line; t ATM is the current line air temperature; t O The air temperature is the freezing point; r RH is the air humidity of the line; v WIND is the wind speed of the line.

7. The line ice thickness estimation method based on stress time series change points according to claim 1, It is characterized in that In step S4, the stress reference value of the line F BM for: (7) in: is the unit deadweight load of the line conductor; is the spacing between line towers; l LE is the length of the line.

8. According to the method for estimating line ice thickness based on stress time series change points in claim 1, It is characterized in that In step S5, whether there is a change point in the ice stress time series is detected based on the Bayesian online change point detection method.

9. A line ice thickness estimation system based on stress time series change points, based on the line ice thickness estimation method based on stress time series change points according to any one of claims 1 to 8, It is characterized in that It includes a data acquisition and processing module, a key factor analysis module, an icing stress time series analysis module and an icing thickness assessment module; the data acquisition and processing module is used to collect multivariate time series data affecting icing, perform data preprocessing, and then send it to the key factor analysis module; The key factor analysis module is used to perform kernel principal component analysis on multivariate time series data, obtain key factors of line icing, calculate the initial value of icing thickness, and determine whether the basic conditions for icing are met; the icing stress time series analysis module is used to calculate the icing stress time series when the basic conditions for icing are met, and detect the change point of the stress time series; The ice thickness assessment module is used to calculate the thickness of ice on the line.

10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the line ice thickness estimation method based on stress time series change points described in any one of claims 1 to 8 is implemented.

Citation Information

Patent Citations

  • Method for predicting transmission line icing thickness

    CN101709959A

  • Method for measuring icing mass of transmission line

    CN101929886A

  • Power transmission line equivalence icing thickness acquisition method

    CN105258664A

  • Power transmission line ice-coating monitoring system

    CN107356218A

  • Power transmission line icing early warning method based on micrometeorological and icing historical data

    CN108921396A

Cited By

  • Power transmission overhead icing multi-source information cooperative monitoring method

    CN121409314A

  • Power transmission overhead ice-coated multi-source information cooperative monitoring method

    CN121409314B